# Abhishek Paatni: everything on this site, as plain text Source: https://paatni.vercel.app/ · updated 15 November 2026 Abhishek Paatni is a Product Designer and Design Engineer based in Delhi, India. Remote preferred; open to on site in Delhi or Bengaluru. He is looking to join established product teams and financially sustainable scaleups solving meaningful problems in AI, B2B SaaS, fintech and healthtech. ## In one line I find the real problem, shape the product, and bring it to market. Product Designer and Design Engineer building AI enabled products and complex workflows, from discovery and interaction design to production ready interfaces. ## By the numbers * 4+ Years developing my craft across design and product * 2300+ Activated mentor profiles, from 0 in 2 months at AceAi Club * 5 Startups I’ve worked with * 21 Days to deploy first functional healthtech product at Vinyasa ## What he does best * Design Strategy: Led product strategy and UX for startups as the first designer, leading project teams of 4 to 7. My drone response system placed 3rd of 100+ teams, then 5th of 20 worldwide, at FlytBase x Design Fusion 2026, a global design competition. * Idea to Market: Took a mental health startup from early Figma frames to a working MVP in 21 days. Designed and built three mobile apps in 2026, already with beta testers. * Visual Craft: Builds design systems and tokens by hand in CSS. Published redesigns of two live SaaS products, Clacky and Seenpaid, each backed by a conversion audit. * Pixel Perfection: I work in the same codebase as the engineers, so what ships matches the design. Automated checks stop a release if any content goes missing. ## Quick answers Q: Who is Abhishek Paatni? A: A Product Designer and Design Engineer in Delhi, India, building AI enabled products and complex workflows: 4+ years and 5 startups, taking products from 0 to 1 and from 1 to 10, often as the founding or first designer. He builds the prototypes and MVPs himself, and keeps one resume for each track. Q: What kind of problems can he own end to end? A: Complex, multi step products, from the first interview to what ships: Onboarding and activation, scheduling, regulated health flows and AI assisted experiences. Reading between the lines of user interviews to find the real problem, the jobs to be done and the riskiest assumptions. Information architecture, interaction design, design systems and edge cases, then shipping with engineers in the repository. At Vinyasa he found why people dropped out of onboarding and shipped the fix in 21 days. Q: Has he worked hands on with AI and engineering? A: Yes. He builds with Claude Code and Codex, driven by written specs and automated tests. Start Anyway: a Flutter app whose Node function plans each day with Gemini, falls back to Groq and validates every reply; about 1,250 tests in CI. Vinyasa 2026: a Next.js and TypeScript app with a privacy first voice analysis API. Q: What evidence shows the impact of his work? A: Vinyasa: more than 75% of new users who started the onboarding questionnaire left before finishing it. His voice check in replaced it; the first functional MVP shipped in 21 days, was chosen for an invite only investor showcase and was later sold. ACEAI CLUB: grown from 0 to 2,300+ activated mentor profiles in 2 months, counting mentors who completed sign up and created a profile. Cryptoventures: the rates advertisers paid rose about 2.7x against the legacy site after his redesign. Tanith Intelligence: 3rd of 100+ teams, 5th of 20 in the global jury round. Pick Your Travels: 33% more qualified enquiries; StageX Lab: 13 international events. Q: Which of his case studies should I read, and what does each show? A: All six, in this order, each with its status: TmAi (Research led test product and interactive prototype): Research, business thinking, complex workflows and conversational UX. Tanith Intelligence (Designathon / competition concept): AI operations, systems thinking and high stakes interface design. Vinyasa (Shipped product; subsequently sold): The strongest evidence of real product problem solving, and healthtech experience. Start Anyway (Independently built mobile app): Independent product ownership and AI assisted product design. unfold (Independently built mobile app): Interaction design, mobile UX and product execution. Menial (Independently built browser game): Design engineering at full depth: story, game systems, pixel art, rendering and sound, built from scratch with no engine. Q: Where can I read or download his resume? A: On paatni.vercel.app/resume/: two one page resumes, Product Designer and Design Engineer, each previewed with its own PDF download. Job titles and dates are the same on both and on this site. Q: Can he work on a product that already exists, with a team in place? A: Yes. He takes products from 0 to 1 and from 1 to 10: ACEAI CLUB: grew the platform from 0 to 2,300+ activated mentor profiles in 2 months. Cryptoventures: as Lead Product Designer he redesigned an established platform with its developers; the rates advertisers paid rose about 2.7x against the legacy site. Vinyasa: after the MVP he designed and built the practitioner platform and the Vinyasa Marketplace. Q: Has he led teams or mentored people? A: Yes. As the founding or first designer he led the project teams of 4 to 7 that grew around his products; project teams, not a design organisation. Tanith Intelligence: led product strategy, information architecture and UX for a team of four; 3rd of 100+ teams, 5th of 20 in the global jury round. Era of No Code: picked by three mentors to build for founders, steered first time founders to shadow test ideas, then mentored cohort students. Taught a Framer workshop to 93 people at IIT Delhi and hosted a meetup there on Figma semantics and design systems. Q: Where is he based, and how does he like to work? A: Delhi, India. He prefers remote work and is open to on site roles in Bengaluru first, then Delhi. Q: How strong is his business acumen? A: He designs with the business model in view, not only the screen: Cryptoventures: native ad placements and an exit intent flow lifted the rates advertisers paid about 2.7x against the legacy site. ACEAI CLUB: an activation journey with payments built in, grown from 0 to 2,300+ activated mentor profiles in 2 months. TmAi: the business model and pricing tiers of an airline concierge subscription. Pick Your Travels: from scratch to its first 133 booked customers in its first 2 months. Era of No Code: steered first time founders to shadow test ideas before paying to build them. Q: How can I contact him? A: Email thepaatni@gmail.com or message him on LinkedIn or X; you can also book a call. Instagram is personal. He prefers remote work and is open to on site roles in Bengaluru first, then Delhi. ## About I’m Abhishek Paatni, a Product Designer and Design Engineer who is a hands on builder and enjoys getting close to a problem before deciding what needs to be made. Over the last few years, I’ve moved across product strategy, interface design, and development. I enjoy working across the entire journey, from asking the first uncomfortable question to seeing a product take shape and become something people can actually use. I’m naturally curious about how things work, enjoy meeting people, and like exploring ideas beyond the boundaries of my job title. Outside work, I have rescued and cared for four dogs for around five years. It’s a meaningful part of my life and a reminder that responsibility is built through patience and consistently showing up. ## Experience * Independent (freelance and contract), Product Designer and Design Engineer, 2021 to 2025: Products, MVPs, design systems and websites for founders in India, the US and the EU, often several engagements at once. * ACEAI CLUB (now Garage Labs), Product Designer and No Code Developer, 2025: From a nonfunctional Airtable prototype to a live platform: activation journey, LinkedIn sign in, payments and Google Calendar sync. Grew it from 0 to 2,300+ activated mentor profiles in 2 months. * Cryptoventures, Lead Product Designer, 2024 to 2025: Rebuilt a legacy, nonresponsive platform as a fully responsive product. Native ad placements lifted the rates advertisers paid about 2.7x against the legacy site (company since closed). * Era of No Code, In House Builder and Mentor, 2024 to 2025: Selected by three mentors to build for founders. Delivered 19 projects (websites, landing pages, MVPs), then mentored cohort students. * Vinyasa, Founding Designer and Builder, 2024: First functional MVP in 21 days from early Figma frames, then the practitioner platform and the Vinyasa Marketplace. Shipped, selected for an investor showcase, and later sold to a US mental health startup. * ConvertML, Designer, 2021 to 2022: Sole designer on a GenAI customer analytics platform: logo, guidelines and the responsive marketing site from scratch. ## Achievements, in his own words * Shipped a healthtech product that went on to be sold (Vinyasa · 2024 · founding designer and builder): I took Vinyasa from early Figma frames to its first functional product in just 21 days. That MVP was picked for an invite only investor showcase and was later sold to a US mental health startup. * Led my team to 3rd of 100+ teams (Tanith Intelligence · FlytBase x Design Fusion Season 3 · 2026): I led product strategy, information architecture and UX for a team of four on an AI native drone command system. We placed 3rd of 100+ teams, then 5th of 20 in the global jury round. * Grew a platform from 0 to 2,300+ activated profiles (ACEAI CLUB · 2025 · product designer and no code developer): I took a nonfunctional Airtable prototype to a live platform, designed the activation journey that turns a sign up into a bookable mentor, and grew it from 0 to 2,300+ activated mentor profiles in 2 months. I take products from 0 to 1 and from 1 to 10. * Lifted ad rates about 2.7x with one redesign (Cryptoventures · 2024 to 2025 · lead product designer): Working with its developers, I rebuilt an established, non responsive platform as a responsive product and replaced intrusive ads with native placements. Advertisers then paid about 2.7 times the rates they paid on the legacy site. * Grew qualified enquiries 33% for a travel company (Pick Your Travels · side project · first digital presence): I built Pick Your Travels’ first digital presence in 7 weeks. It brought 33% more qualified enquiries, a 4.7% visitor to lead conversion and 17+ DMC partnership applications a quarter. * Invited to teach 93 people at IIT Delhi (IIT Delhi · official Framer workshop · invited speaker): I was an invited speaker at an official Framer workshop for 93 attendees, and I also hosted a community meetup at IIT Delhi on Figma semantics and design systems. * Hand picked by three mentors to build for founders (Era of No Code · 2024 to 2025 · in house builder and mentor): Three mentors chose me from the cohort to build for founders across the network. Over 19 projects I steered first time founders to shadow test an idea before paying to build it, and I was brought back to mentor cohort students. * Won the founder over with evidence, not opinion (Vinyasa · 2024 · research and stakeholder decision): Almost every comparable platform used a long form, and more than 75% of our new users left ours unfinished. I read therapy transcripts, interviewed people in therapy, showed the founder what the form was costing and replaced it with a 2 to 3 minute voice check in. * Kept users’ health data where it belonged (Vinyasa · 2024 · decision with engineering): When a shortcut would have moved voice recordings off the platform, I said no, built the audio conversion middleware myself and set HIPAA aligned data handling for US users. We still shipped in 21 days. * Built a whole game with no engine (Menial · 2026 · solo): I designed and built Menial alone: a grimdark pixel art shooter with its own WebGL2 renderer, my own sprite tools and every sound synthesized in the browser. Chapter I is playable now. ## Selected work, in his order 1. TmAi, Flight Concierge (Research led test product and interactive prototype): Research, business thinking, complex workflows and conversational UX. 2. Tanith Intelligence (Designathon / competition concept): AI operations, systems thinking and high stakes interface design. 3. Vinyasa (Shipped product; subsequently sold): The strongest evidence of real product problem solving, and healthtech experience. 4. Start Anyway (Independently built mobile app): Independent product ownership and AI assisted product design. 5. unfold (Independently built mobile app): Interaction design, mobile UX and product execution. 6. Menial (Independently built browser game): Design engineering at full depth: story, game systems, pixel art, rendering and sound, built from scratch with no engine. Together the six show a product designer who starts from research, thinks in business models and systems, has shipped a real healthtech product that was later sold, and designs and builds AI enabled products, mobile apps and even a browser game end to end himself. ### TmAi, Flight Concierge Status: Research led test product and interactive prototype. What it shows: Research, business thinking, complex workflows and conversational UX. TmAi tests one question for an airline: could it sell a seat without paying somebody else to sell it? The answer reshaped the project from a booking app into a distribution strategy. * Project: TmAi, an airline concierge subscription with a conversational booking agent * Client: Rebecca Doshi, who brought the idea, Bangalore * Role: Research, product design, prototype * Year and timeline: 2024, about four months * Deliverables: Research, unit economics, product strategy, a three tier membership, Figma designs and a working conversational prototype * Status: A case study and prototype. No flight is booked and no payment is taken #### The brief Rebecca Doshi brought the idea fully formed and asked whether it could hold up: could an airline sell a seat without paying somebody else to sell it? The finding that reshaped the project: the problem was never the seat. It was the way the seat gets sold. #### The problem, through one frequent flyer Abhishek sat with eleven people like him and combined them into one persona, Arjun, a sales lead who flies four or five times a month, mostly Bangalore to Delhi, usually booking two days out. * Seventeen tabs, and he is still not sure of the fare. * At quarter end he digs four months of confirmation emails out of his inbox to file receipts. * No airline app lets his colleagues pay for their own seats on the same trip. #### The research IATA put the industry at 979 billion dollars of revenue in 2025 and 36 billion of profit: a 3.7 percent margin, around 7.20 dollars per passenger per flight. Roughly half of all tickets still sell through a travel agent or booking system, each costing the airline about eight percent. On 489.5 billion dollars of indirect sales that is 39.16 billion dollars a year, more than the industry’s profit. That moment turned the project from a booking app into something about distribution. #### The solution: a conversation instead of a form A normal return booking is six dropdowns, a results page, a fare table, a seat map and a payment sheet. TmAi replaces it with a concierge agent inside the airline’s own app: one button hands over the trip, the agent searches, compares, shows three to five options and books the one you pick. Colleagues are tagged with an @ and receive a payment request for their own seat, and seats beside the traveller are held for thirty minutes at the same fare. The receipt is filed for expenses. #### Pricing An annual membership in three tiers named after cabins, so the airline earns from a passenger even in months they never fly. * Economy Plus, 100 dollars a year: books on saved preferences and moves the booking onto the airline’s own channel. * Business Class, 700 dollars a year: priced against risk for people who fly weekly. Predicts delays and rebooks early, refunds a fare that drops, checks visas, turns a quarter of receipts into one expense report. * First Class, 2,000 dollars a year: attention as the product. A person on the line in under a minute, any seat without a fee, lounge access and held upgrades. #### How to read the numbers The 3.7 percent to 11.7 percent margin is modelled from public industry data (IATA 2025, T2RL 2022) and the assumptions written in the case study. It is not a measured result from any carrier. Selling direct keeps the eight points that distribution takes: 3.70 dollars becomes 11.70 dollars on a hundred dollars of ticket. Links: Read the full case study and try the prototype: https://tmai-casestudy.vercel.app/ ### Tanith Intelligence Status: Designathon / competition concept. What it shows: AI operations, systems thinking and high stakes interface design. Tanith is an AI native command system for critical airspace, designed for FlytBase x Design Fusion Season 3. Its premise: AI is the agent and the human is the judge. * Project: Tanith Intelligence, an AI native command system for drone based disaster response * Context: FlytBase x Design Fusion Season 3, 2026 (a design competition) * Role: Led product strategy, information architecture and UX for a team of four; built and shipped the case study website solo in Framer * Timeline: 9 days from concept to shipped * Result: 3rd of 100+ teams in the qualification round; 5th of 20 in the global jury round * Status: A competition stage prototype and concept, not a shipped product #### The brief The brief named a 48 minute gap between a disaster being spotted and a coordinated response reaching the ground, and asked for 90 seconds. In the 2018 wildfire at Mati, Greece, 104 people died, most of them in the first 90 minutes. Detection is already fast. Coordination is not. #### The reframe Every other team redesigned the dashboard. Abhishek redesigned what the operator is for: AI is the agent, and the human is the judge. #### What was built Tanith is an AI native command system for critical airspace. The AI detects, verifies and precoordinates the whole response. The human approves anything irreversible. It has two operating states: Idle overwatch, where silence is the safety signal, and a Focused alert view that collapses the screen to one decision. Every AI call, operator action and failsafe is logged to a legal audit layer. #### How the 90 seconds works * T+0s: a sensor trips. Three AI layers clear the noise. The quickest drone launches, not the closest. * T+30s: the drone is in transit. Four feeds spin up: night vision, thermal, 4K, and a map with trajectory. * T+60s: the AI assesses the feeds against historical wildfire patterns and live weather. Confidence climbs. * T+90s: wildfire confirmed at 94%. The operator proceeds or aborts, or a 25 second failsafe proceeds and logs it. Ground teams are already moving. #### Four decisions, and what each one cost * Human as judge over full autonomy. Cost: seconds of latency on every irreversible action. * A new enterprise identity instead of the rebrand the brief asked for. Cost: jury points. * A Dead Man’s Switch with four escalation layers, not one. Cost: complexity. * Quickest not closest dispatch. Cost: explainability. #### What Abhishek owned Product strategy, information architecture, UX, the 90 second decision chain, and the live case study site, written, designed, built and shipped solo in Framer. He led a team of four. #### Result and feedback Tanith placed 3rd of 100+ teams in the qualification round and 5th of 20 in the global jury round of Design Fusion Season 3. Prathmesh, Lead Product Designer at FlytBase, wrote: “You people have gone way outside the problem statement and truly tried to work out a solution that can actually be called an AI native solution with human in the loop.” #### How to read the numbers The 90 second response is the target set by the competition brief, shown working in the prototype. It is not a measured field result. Tanith is a competition stage prototype with a research foundation and a clear architecture, not a finished product. Links: Read the full case study: https://tanith-intelligence.vercel.app/ ### Vinyasa Status: Shipped product; subsequently sold. What it shows: The strongest evidence of real product problem solving, and healthtech experience. Vinyasa helped people reach the right healthcare professionals, including therapists and psychiatrists. More than 75% of new users dropped out of its onboarding form. Abhishek found out why, then designed and built the fix end to end in 21 days. * Client: Vinyasa, founder Divya Shah * Role: Founding Designer and Builder: Product Designer and In house Builder * Original MVP: 2024, 21 days from early Figma frames to a functional MVP * Rebuild: 2026: the same idea rebuilt from scratch in five days on free tools, as a free progressive web app with no account * Platform: Progressive web app (rebuild) * Outcome: The original MVP was selected for an invite only investor showcase and was later acquired by a US mental health startup #### The brief and the problem Every new user first had to complete a long onboarding questionnaire, and too many never did. More than 75% of new users dropped off during onboarding, leaving them without help, therapists unable to support them and the business struggling to move forward. Long forms cost people who are already in distress the most. Across 93 mental health apps, about 4 in 100 users were still active after 15 days (Baumel et al., 2019). #### How the answer was found Abhishek read therapy session transcripts, then spoke one on one over video with people going through therapy on Vinyasa. The transcripts showed what people said. The calls showed how they felt: their voices carried emotion the written words did not. He understood why the team used a long form, since almost every comparable platform does, and convinced the founder to try a different approach. #### Three insights that shaped the design * Long forms ask too much of people who are already struggling. Fix: short prompts, answers you tap, and spoken tasks instead of an essay style form. * Voices carry what words miss. Fix: collect emotional signal from the voice instead of relying only on forms. Voice patterns are a signal, not a verdict. * Trust has to be designed. Fix: plain language scores, a clear “not a diagnosis” label, and no account or stored history. #### The solution: a short check in that listens Vinyasa became a short mental wellness check in that uses a person’s own voice to understand how they are doing, with a few simple tasks instead of a long questionnaire. In four steps: check in, reflect, listen to a sound chosen for what you need most, and recheck how you feel. Scores are self reported and labelled “not a diagnosis”. If several areas run very high, it points to a doctor, a therapist or a helpline. #### The original MVP, 2024 Abhishek joined when only early Figma frames existed and shipped the first functional MVP in 21 days: a 2 to 3 minute self assessment using audio, visual and questionnaire input to route people to the right type of mental health professional. The MVP used the Thymia API to read emotional information from voice. He rejected an off platform workaround for converting audio (WAV vs MP3) because it would move sensitive health data off the platform, and built the secure on platform conversion middleware himself, with HIPAA aligned data handling for US users. #### More at Vinyasa * The practitioner platform, designed and built over roughly three months: scheduling, client notes and records, conflict detection, double booking prevention, hold and release on unconfirmed bookings and cancellations. * Vinyasa Marketplace (live at theravision.vinyasahealth.com), for therapists and psychiatrists: providers are onboarded only after a valid medical licence number is verified, and consumers are onboarded by the internal team to prevent spam and abuse. #### The 2026 rebuild Two years on, Abhishek rebuilt the idea from scratch with the skills he has now and made it free to try. It is a Next.js and TypeScript progressive web app with a privacy first voice analysis API, 22 soundscapes synthesised in the browser with Web Audio, and Playwright end to end tests that assert nothing is stored and no third party host is contacted. It went live on 15 October 2026. Because it runs on free tools, its voice readings are less accurate than the original’s. It does not claim to fix mental health issues or act as a treatment. #### What he took from it A check in only works if it asks very little. A product only works if people trust it. Both came from listening before designing anything. Links: Read the full case study: https://vinyasa-casestudy.vercel.app/ · Try the 2026 rebuild: https://vinyasa-app.vercel.app/ · Vinyasa Marketplace: https://theravision.vinyasahealth.com/ ### Start Anyway Status: Independently built mobile app. What it shows: Independent product ownership and AI assisted product design. Start Anyway helps you reach big goals by turning them into small daily tasks, with a personal companion that guides you every day and holds you accountable. * Product: Start Anyway, a free Android app: no account, no ads * Role: Founder. Designed the progression loop and the identity, and directed the build * Status: Android beta, version 1.0.0 beta 3, released 2 November 2026 (20.5 MB). Not on Google Play yet, so it installs from a file * Stack: Flutter and Dart, a Node serverless function, offline first with a local database * Quality: About 1,250 automated tests (Flutter and Node) in CI on GitHub Actions, a signed release and APK checksum verification #### Why it exists Abhishek is a power user of productivity apps, and they all had one thing in common: he would spend 30 minutes, sometimes an hour, writing, changing and editing to dos, and still prioritise the wrong tasks because he had forgotten the bigger picture while making the list. #### How it works * Tell your companion two goals, by typing or speaking. It shows what it understood, for you to confirm. * Pick the plan you can keep: All In (120 minutes a day), Real Life (50) or Tough Season (20), made from what you said about your life and what usually stops you. * Choose today’s time (15, 30 or 60 minutes). The plan fits the time you have: two to four tasks, never a pile. * Press and hold Start for about a second, and the task begins. Leave any time, and your companion asks first. #### The companion You choose how your companion looks and give it a name. There are four, each with its own voice and movement: Pointer, Up, Navigator and Floating Parts. #### Every day adds up Finish a task, finish part of one, or focus for five minutes, and the day counts. Your wall fills in poster by poster. A quiet day stays a blank poster, never a red mark, and the next day starts fresh. Distraction Reminders are opt in and off until you switch them on. The app only knows which app is open and for how long. #### Data and privacy Goals, plans and history stay on the phone. To plan a day, the app sends your goal text and a short summary of your last seven days to an AI service: Google Gemini on its free tier, or Groq when Gemini is not available. The AI function validates every response against a shared schema and applies rate limits and a kill switch. Links: Visit Start Anyway and download the beta: https://startanyway.vercel.app/ ### unfold Status: Independently built mobile app. What it shows: Interaction design, mobile UX and product execution. unfold is a free two player question game for one phone. Pass it, answer out loud, and skip the small talk. * Product: unfold, a free Android game for two players on one phone * Role: Product designer and builder * Status: Android version 2.4, 10 MB, Android 10 or newer. Released 28 October 2026. Not on the Play Store or App Store yet, so it installs from the site * Privacy: No account, no internet permission (it works offline), and the only permission it asks for is vibration. Names and answers vanish when you close it #### The idea Small talk is mid. unfold turns the awkward first minute into a reason to connect: two people, one phone, real questions, answered out loud. #### The whole game, in seven screens * Name the players, and pick how the app hypes you. * Set the vibe, from chill to real, flirty or unhinged. The whole screen changes mood. Unhinged is 18+ and both players confirm before it opens. * Roll the die: one to six rounds each. On a phone you can shake it. * Answer out loud, one question at a time. * Ate or veto: swipe the card right or left and the app reacts. Aura is ate minus vetoes, and the highest aura wins. * Crown a winner. * Get the receipt: the winner takes a selfie and unfold turns it into a story poster with the score. Nothing is saved unless you say so. #### Design decisions The app never asks for internet permission, so it works on a plane, and it keeps nothing about the players. It only remembers sound, haptics and calm motion settings, and that you have seen the intro. Links: Visit unfold and download it: https://unfoldnow.vercel.app/ ### Menial Status: Independently built browser game. What it shows: Design engineering at full depth: story, game systems, pixel art, rendering and sound, built from scratch with no engine. Menial is a grimdark 2D action shooter for desktop browsers, set in July 1919: the Great War has ended, the sky has opened, and humanity is still fighting itself. Story, systems, art, rendering and sound are all built from scratch, with no game engine and no libraries. * Product: Menial, a 2D pixel art action shooter for desktop browsers * Maker: Abhishek Paatni, working alone * Status: Chapter I, Peace Without Bread, is playable in desktop browsers with a keyboard and mouse: an opening cinematic, champion select, three faction prologues and the chapter itself * Built with: About 6,400 lines of plain JavaScript and WebGL2 with no libraries, sprites made by his own Python tools, and every sound synthesized in the browser with Web Audio * Platforms: Desktop Chrome, Edge, Firefox and Safari with WebGL2 #### The game You choose one of three champions: Anja Rook of the Order, Dr. Elias Venn of the Directorate or Kade Vey of the Compact. Each faction arms its champion in its own house, teaches the weapons in a prologue and sends them to the front its own way. Chapter I, Peace Without Bread, puts the three human factions on one front with the Veyr, and builds through a duel with a rival champion and a Warden to the Mouth in the Mud, a creature that hunts by sound. #### Design decisions * The rival champion is played by an AI that drives the same controller as the player, so it moves, aims and reloads under the same rules. * Crouching behind sandbags means only plunging fire can reach you, and crouch walking is silent, which matters against a boss that hunts by sound. * A reload takes at most a second and a weapon swap takes 40% of a reload. The field map does not pause the game. #### How it is built * The game renders at 640 by 360 in WebGL2, with multiple render targets for lighting, puddle reflections, bloom, shockwave distortion and film grading. * Characters are hand drawn pixel sprites made by his own Python tools: body parts drawn as pixel spans, legs posed from CMU motion capture clips. The arms and weapon are a separate sprite that turns at the shoulder to follow the aim. * Environments are painted in code when the game loads, and every sound, music state and ambience is synthesized with Web Audio. #### Where it stands Chapter I is playable now in a desktop browser, free, with no install and no sign up. It needs a keyboard and mouse. Links: Play Menial: https://menial-ten.vercel.app/ ## Side projects (Framer case studies) ### Clacky: Redesigned Clacky’s Homepage Around One Promise, on a Design System Built From Scratch. A conversion audit of clacky.app, a playful Mac app by solo maker Justinas Rolando, followed by a new hero and design system that put one promise, the real product and the trial terms in the first screen. * Product: Clacky, a Mac app by Justinas Rolando with click sounds, a screen recorder, Mac utilities, a cozy town game, analytics and a post scheduler (as audited) * Role: Strategist, Designer and Developer * Timeline: Audit 19 Oct 2026; brief, hero and design system 20 Oct 2026 * Project type: Conversion audit and hero redesign * Built with: Plain HTML, CSS and a little JavaScript, with no framework or UI library, deployed on Vercel * Deliverables: A 4 page conversion audit, a copy and visual design brief for the whole homepage, a live hero section and its design system * Built to move: +15 to 25% hero CTA click through; 10 to 20% fewer checkout drop offs; +20 to 30% CTA taps on phones, against the page as it was on 19 Oct 2026 * Note: clacky.app has since changed its offer, so the hero’s trial and price copy reflects the page as it was on 19 Oct 2026 #### Summary Clacky launched on 28 September 2026 as a Mac app with click sounds, a screen recorder, notch tools, a cozy town game, analytics and a post scheduler. People loved its playful website, but by mid October new trials had slowed to a trickle. Abhishek read the page at desktop and phone width, the trial page and about 220 of Justinas’s posts, wrote a conversion audit, then turned its top fixes into a copy and design brief and a new hero with its own design system. #### The problem * The promise said everything. The headline covered “everything you make”, and the subhead listed sounds, a recorder, 20+ tools, analytics and a harbour town. * No product in the first screen. Desktop showed an illustration; on phones the subhead was hidden, so the hero was a headline and a button. * The card requirement was hidden. The button said “7 days free · cancel anytime”; the card appeared only on the trial page and in FAQ 8 of 12. * The page was long and leaky: about 15,000 px on desktop and 28 screens on a phone, with the next button after the hero some 15 screens down. * Deciding facts came late. Pricing sat before the competitor comparison, and the macOS requirement was in the FAQ, 82% down the page. #### What he changed He gave the hero one story for one kind of visitor, makers who build and share from a Mac: make it, show it, see what works. The real product sits in the first screen, and what a visitor needs to decide sits beside the button. #### Key moves * A headline that ties the features into one workflow: “Make it. Show it. See what works.” * A Studio window that turns a screen recording into a captioned, ready to share demo, playing once and ending on its final frame. * A small “Task completed” notification and an analytics panel labelled “Sample data” as supporting proof. * Trial terms beside the button: “Card at checkout, $0 today. Cancel anytime.”, every feature included, 2 Macs, macOS 13 or later. * The trust strip kept right under the hero: notarized by Apple, recordings stay on your Mac, cookie free analytics. * A shorter navigation (Features, For makers, Pricing, Privacy) and, on phones, a menu, a header button and the full explanation. * A design system page with every token and component, where each accent colour has one job. * Visitors who prefer reduced motion see the finished state straight away. #### The result A live hero and design system at clacky-hero.vercel.app, plus a brief with copy for all 17 homepage sections. It is built for 15 to 25% more hero clicks, 10 to 20% fewer checkout drop offs and 20 to 30% more taps on phones, against the 19 Oct page. Links: See the Clacky redesign: https://clacky-hero.vercel.app/ · See the design system: https://clacky-hero.vercel.app/designsystem ### Tanith website: Built the Website for a Drone Concept Ranked 3rd of 100+. A Framer website that presents Tanith Intelligence, his team’s AI native drone response concept, the way an emergency service would meet it: what it does, how the 90 seconds work, and who stays in charge. * Project: Website for Tanith Intelligence, a concept for drone based disaster response * Context: FlytBase x Design Fusion Season 3, 2026 (a design competition) * Role: Designer and Framer Developer, in a team of four * Built with: Framer, with desktop and mobile layouts * Result: The Tanith concept placed 3rd of 100+ teams in qualification and 5th of 20 in the global jury round * Note: Tanith is a competition concept, not a shipped product. The 90 second response is the brief’s target, not a measured result #### Summary Tanith Intelligence is an AI native command system concept for drone based disaster response, made by a team of four for FlytBase x Design Fusion Season 3. The full case study is in Selected Work; this is the product website around it. The site walks a city or emergency service through Tanith in order: the promise, a three step chain of command, the operator’s view, a 30 day rollout and a way to talk to the team. #### The problem * A complex system. Sensors, drones, AI layers and human operators had to read as one simple story. * Fear of autonomy. Emergency teams need to see that a person still makes the irreversible calls. * A bold promise. “Under 90 seconds” had to be explained step by step, not just claimed. * A careful buyer. Cities and agencies want to know how deployment and training work before they get in touch. #### What he changed He turned the concept into a website that opens on the moment that matters, “When Every Second Counts, Tanith Responds.”, then explains how the response works with the human in charge at every step. #### Key moves * Opened on a wildfire at dusk with one line of promise and two actions: Talk to Our Team and See How It Works. * Explained the response as a chain of command: Detect, Confirm, Act. * Showed the operator’s screens under “Operator Centered, Not AI Obsessed”. * Answered the trust question directly: operators can override, confirm or question the AI, and every action is logged. * Turned adoption into a 30 day plan: site audit and zone mapping, operator training and drills, then live deployment. * Closed on “Let’s Make 90 Seconds the New Standard.” with a direct line to the team. * Used monospaced display type and a single violet accent for a command room feel, on desktop and mobile. #### The result A live Framer site that gives the Tanith concept a public face, from the first promise to a request to talk. The concept behind it placed 3rd of 100+ teams in qualification and 5th of 20 in the global jury round. The 90 second response is the target the site is built around, not a measured result. Links: Visit the Tanith website: https://tanith.framer.website/ · Read the Tanith case study: https://tanith-intelligence.vercel.app/ ### Seenpaid: Repositioned Seenpaid Around Revenue per Post, on a Design System Built From Scratch. A conversion review of seenpaid.com, Justinas Rolando’s social media scheduler, and a new hero that leads with what makes it different: seeing which posts bring in money. * Product: Seenpaid, a social media scheduler by Justinas Rolando that AI agents can drive and that connects tracked clicks to payments * Role: Strategist, Designer and Developer * Timeline: Review 19 Oct 2026; brief, hero and design system 20 Oct 2026 * Project type: Conversion review and hero redesign * Built with: Plain HTML, CSS and JavaScript on Vercel, with a small Node script that builds the design system page from the hero’s own markup * Deliverables: A 3 page conversion review, a copy and visual design brief for the whole landing page, a live hero section and its design system * Built to move: +15 to 25% hero CTA click through; 10 to 20% fewer drop offs at the card step; 10 to 15% lower hero bounce rate, against the page as it was on 19 Oct 2026 * Note: Figures in the hero are sample data and are labelled that way #### Summary Seenpaid lets you write a post once, publish it across social platforms yourself or through an AI agent, and see which post brought in a sale. The page was polished and honest, but it led with a crowded category and surfaced its costs late. Abhishek reviewed the site on desktop and mobile, the pricing and signup pages, and Justinas’s posts about Seenpaid, then wrote a copy and design brief and built a new hero with its own design system. #### The problem * The hero sold the commodity. “Let your AI agent run your socials” put it beside Buffer and Hootsuite; revenue per post, the real difference, was a tagline. * The message flipped along the way: the title tag and signup page led with revenue, the hero with AI scheduling. * Costs surfaced late. The card requirement lived in the FAQ, and posting to X needs your own paid X API key. * “21 platforms” overstated what was live: Instagram, Facebook and Pinterest were marked SOON. * Jargon up top. An MCP badge and agent language could lose a nontechnical social media manager. * No social proof: no testimonials, customer count, case study or reviews anywhere on the page. #### What he changed He rebuilt the hero around one piece of content moving from post to payment, so a visitor sees the whole promise in one look and reads the costs before they click. #### Key moves * Led with the difference: “Stop guessing which posts make you money.”, which now matches the title tag and the signup page. * Showed one post travelling: a calendar with it published on four channels, the AI agent card that scheduled it, and a results board with matched sales. * Kept attribution honest on screen: matched and unmatched payments are shown apart, and every figure is labelled sample data. * Put the terms beside the button: “7 days free. Card required. $0 due today.”, cancel before day 8, and posting to X uses your own X API key. * Spoke plainly, with “your AI agent” instead of protocol names and a four link navigation: Product, Platforms, Pricing, Resources. * Kept the founder in the frame with a small Justinas cursor on the results board. * Designed mobile as its own layout: a trial button in the header, one results panel and a four step cue from Scheduled to Payment matched. * Generated the design system page from the hero’s own markup, with one accent colour and every token listed. #### The result A live hero and design system at seenpaid-hero.vercel.app, plus a brief with copy for the full landing page, including a platform grid that marks what is live, limited or coming soon. The hero fixes the message and the cost surprises. Real social proof is still the biggest gap, and it needs customers, not design. It is built for 15 to 25% more hero clicks, 10 to 20% fewer drop offs at the card step and a 10 to 15% lower bounce rate, against the 19 Oct page. Links: See the Seenpaid redesign: https://seenpaid-hero.vercel.app/ · See the design system: https://seenpaid-hero.vercel.app/designsystem · See the old Seenpaid design: https://seenpaid.com/ ### StageX Lab: Built StageX Lab’s Digital Presence That Secured 13 International Events. He established the first digital presence for StageX Lab, which led to 13 international event opportunities and much stronger credibility with artists, promoters and agencies. * Client: Rajat Srivastav, founder of StageX Lab * Role: Strategist, Designer and Framer Developer * Timeline: 4 weeks * Project type: GTM and marketing website * Results: 13 international events secured; 75% average scroll depth; 6 in 10 visitors inquired * Note: The StageX Lab website is not reachable at the moment, so there is no live link here #### Summary StageX Lab built its reputation through referrals and industry connections, managing high energy events from Holi festivals and techno experiences to celebrity concerts. Despite executing large scale productions, the agency had no digital presence to show what it could do. Abhishek partnered with the founder to create the company’s first digital identity, from strategy and UI design to Framer development. The website was built to put their work at the centre of the experience. Instead of generic agency messaging, every interaction celebrates the scale, energy and behind the scenes craft that define StageX Lab. The result was a portfolio first website that strengthened trust with international organisers and helped the agency secure 13 international event opportunities. #### The problem * No digital presence despite managing large scale events. * Heavy dependence on referrals and personal connections for new business. * The agency needed to stand apart from conventional event agency websites. * It needed a platform that showed execution, energy and behind the scenes expertise. #### The solution Instead of designing another agency website filled with marketing copy, he built an experience where the work itself is the hero. Immersive visuals, motion, bold typography and storytelling reinforce the identity, and finished productions sit alongside behind the scenes content so visitors see both the final events and the operational excellence behind them. He also designed a new logo mark for a memorable, scalable identity. Rather than selling services, the website demonstrates capability through the quality and energy of the work itself. * Positioning strategy and logo mark design * Information architecture and website UX * Visual design, Framer development and motion interactions * Portfolio storytelling, mobile responsiveness and performance optimisation #### Design challenges * Avoiding industry clichés. Most event websites look interchangeable, and this one needed to be unmistakably StageX Lab. * Letting visuals tell the story. The work had to communicate credibility without excessive marketing copy. * Motion driven interactions. Animations and transitions reflect the energy of live events. #### Impact * Secured 13 international event opportunities. * Increased digital credibility with global organisers. * Established StageX Lab’s first professional online presence. * Achieved an average 75% scroll depth, with 6 in 10 visitors inquiring. * Created a scalable foundation for future growth. ### Pick Your Travels: Built a Travel Brand From Scratch, With Its First 133 Booked Customers in 2 Months. Delivered a conversion focused website designed to support a 35% increase in qualified enquiries, a 4.8% visitor to lead conversion rate and 25+ high quality DMC partnership applications, through clearly defined user journeys and trust first messaging. * Client: Ansh Agrawal, Pick Your Travels * Role: Designer, Developer and Strategist * Timeline: 7 weeks * Project type: Website and positioning strategy * Built with: Framer, with reusable components, a scalable design system and a CMS for destinations, blogs and packages * Results: First 133 booked customers in the first 2 months; 33% more qualified enquiries; 4.7% visitor to lead conversion; 17+ DMC partnership applications per quarter * Targets: 35% more qualified enquiries; 4.8% visitor to lead conversion; 25+ DMC partnership applications #### Summary Pick Your Travels was not looking for another travel website. They needed a digital foundation that could position them as a trusted travel partner for both Indian travellers and international DMCs. Abhishek partnered with the founders to define the company’s positioning, identify its competitive advantage, research customer behaviour, structure the information architecture and translate those insights into a premium digital experience: a strategy led website built to build credibility, simplify decision making and support long term growth. #### The problem * No digital presence. The business relied entirely on WhatsApp conversations and referrals, so new customers struggled to discover or trust the brand. * Trust was not visible. Despite exceptional travel experiences, there was no platform for testimonials, curated destinations or credibility. * Two audiences, one process. Travellers and international DMC partners had completely different needs, with no dedicated experience for either. * Growth could not scale. Without a structured website, content strategy or lead generation system, expanding into new destinations depended on manual outreach. #### What he changed He built Pick Your Travels’ first digital presence by turning a WhatsApp first travel business into a premium brand with a scalable acquisition and partnership platform. Because the end user is someone opening the website on a phone, the experience stays seamless and premium from desktop to mobile. #### Key moves * Defined the positioning around trust, transparency and premium service. * Designed separate journeys for travellers and international DMC partners. * Established the company’s first digital sales funnel beyond WhatsApp enquiries. * Structured destination, package and partnership content into an intuitive information architecture. * Built a scalable CMS for destinations, travel blogs and curated packages. * Introduced premium visual storytelling to reinforce credibility and quality. * Planned the launch strategy to support organic growth and future expansion. * Developed the complete website in Framer with reusable components and a scalable design system. #### The result A minimal luxury website focused on trust, clarity and conversion. The information architecture separates traveller and DMC journeys while highlighting curated destinations, premium travel packages, customer stories and partnership opportunities. Strategic calls to action, destination focused content, dedicated enquiry flows and a scalable CMS support long term growth. Links: Visit pickyourtravels.com: https://www.pickyourtravels.com/ ### Swish proposal for MK Agency: Turning a Static Pitch Deck into an Interactive Product Proposal for Southeast Asia Basketball. He replaced a traditional PDF proposal with an interactive web experience that communicated product strategy, business vision and implementation roadmap far more engagingly for MK Agency. * Client: MK Agency * Role: Designer, Strategist and Framer Developer * Timeline: 3 days * Project type: Interactive proposal landing page * Results: Secured a stakeholder meeting after previous PDF proposals consistently went unanswered; 24 strategic sections; 3 primary stakeholder journeys #### Summary MK Agency was preparing to pitch a new digital ecosystem to a Southeast Asian basketball organisation. Previous PDF proposals received little engagement, making it hard to communicate the scale of the vision. Instead of another presentation, Abhishek proposed designing and developing the proposal as an interactive landing page, combining product strategy, storytelling, UX thinking, business rationale, implementation phases and visual design in a single experience stakeholders could easily open and explore. The proposal positioned the platform as the digital infrastructure connecting parents, coaches, league administrators and federation officials through one unified ecosystem. #### The problem * Traditional PDF proposals generated low engagement and failed to communicate the scale of the product vision. * Parents, coaches, league managers and administrators needed completely different workflows that were hard to explain in static documents. * The agency needed to communicate product strategy, market opportunity, implementation roadmap and business value before writing production code. * The proposal had to feel premium enough for executives while staying simple enough to understand in one sitting. #### The solution He designed and built an interactive proposal that works like a landing page rather than a document, turning dozens of disconnected slides into one cohesive experience built on a centralised basketball infrastructure, phased pilot modules and multiple stakeholder roles. Instead of treating the project as a presentation, he treated it as a product experience: a scroll based narrative where every section answers the next business question, so that every section builds confidence toward the final proposal. * Current operational problems and market opportunity * Why existing workflows fail * Platform architecture and user roles * Feature scope and pilot rollout * Investment and implementation roadmap Links: View the live proposal: https://swish-proposal.vercel.app/ ## Testimonials (verbatim excerpts) Divya Shah, Former Founder · Vinyasa, worked with him on the Vinyasa MVP, 2024: “Paatni took our product from idea to prototype in 21 days, handling design, research, flows, and development end to end. He doesn’t just make things look good, he makes them work.” Sumit Kumar Singh, Ex Microsoft · Harvard Business School, worked with him on their no code SaaS platform: “He built our no code SaaS platform with speed, precision, and creativity. Abhishek consistently solved complex problems and exceeded expectations.” Sarthak Tiwari, Founder · KaartHQ · Ex IBM: “Paatni is a rare blend of sharp design sense and tech skills. His deep UX understanding and team first mindset make him someone you want on your side not your competitor’s.” Peehu Bansal, Former Founder · Vinyasa, worked with him on 4 Vinyasa projects: “Abhishek helped us bring our product to life using NoCode and AI tools, combining speed with smart problem solving. He built complex logic in Bubble, integrated AI features seamlessly, and approached every challenge with creativity and care.” Aditya Pati, Former Founder · Era of No Code, worked with him on MVPs and websites for founders at Era of No Code, 2024 to 2025: “Abhishek blends sharp product design instincts with no code speed. For any founder looking to develop or ideate MVPs efficiently, he’s a go to builder you can trust.” Komal Patel, Cofounder · Debound, worked with him on Debound’s new website: “Abhishek played a key role in designing Debound’s new website, showing exceptional dedication, creativity, and adaptability throughout the process. He was detail oriented, took ownership of the project, and handled constant feedback with patience.” Aayush Chandna, Founder · Recykal: “Abhishek aka Paatni, is the complete package, a product designer and no code developer who delivers exceptional work with unparalleled agility. His ability to quickly grasp a problem’s root cause and craft creative solutions that is aimed at it is remarkable.” ## Events and community * itsdesignare event: With the founder of itsdesignare, one of India’s biggest design communities, and a good friend. * With Ansh Mehra: The day the Current Edge School launch I worked on for Ansh Mehra went live. * Voxago community head: Hosting a meetup at IIT Delhi on Figma semantics and design systems. * Guest speaker at IIT Delhi: A hands on Framer workshop I taught, with 93 people in the room. * Running a Claude Code session: With the participants after a hands on Claude Code session I ran. * AI Impact Summit: Six hours of AI products, demos and talks at the summit. * Working remotely: Waiting for a client to join a Google Meet from a café. I love working remotely. * FITT IIT Delhi 2026 expo: Invited as a guest to the FITT expo at IIT Delhi. * An electric Austin Seven: At IIT Delhi’s Vintage to Future car expo: an Austin Seven converted to electric in the mid 20th century, after World War II. * With the Claude Code India community head: I spoke at his Claude Code workshops and helped organise them. * Made a new friend at a meetup: At a Claude Code meetup. The friendship started with what they had built and what they hadn’t yet, more than with my own projects. * A Web3 builder at a meetup: A sharp pitch, then a long talk about how crypto adoption could change as AI gets more capable. * Accenture AI expo: Invited to the Accenture AI expo, an invite only event for product designers and builders. * Volkswagen Beetle: A nostalgic modern classic Beetle, and a first generation Type 1C model at that, from the facelift era. * Audi, Dressed as an RS 6: A red C7 generation Audi wearing the face of the 2015 RS 6 facelift: black honeycomb grille, quattro lettering in the centre intake and sharp LED eyes. * Ford Mustang GT: A sixth generation Mustang in Race Red with twin black stripes, the 5.0 litre V8 fastback Ford brought to India in 2016. * Bangla Sahib: I just could not help but capture the beauty of the architecture from the best angle possible. * Selective sweet tooth: My favourite flavour, and the only sweet I can eat all day. * Living independently: Not my own home yet, but a rented place of my choice, to live independently. * Late night call: Helping the junior designer from my team with her task, so she could aura farm in tomorrow’s meeting. ## Resumes Product Designer resume (https://paatni.vercel.app/Abhishek-Paatni-Product-Designer-Resume.pdf) SUMMARY Product Designer with 4+ years across 5 startups in AI, B2B SaaS, fintech and healthtech, taking products from 0 to 1 and from 1 to 10: I grew ACEAI CLUB from 0 to 2,300+ activated mentor profiles in 2 months, and lifted ad rates about 2.7x on an established platform at Cryptoventures. Often the founding or first designer, leading project teams of 4 to 7. I own discovery, user research, product strategy, information architecture, interaction design and design systems, and tie each decision to a business outcome. Strongest in complex workflows: onboarding, scheduling, regulated health data and human in the loop AI. I prototype and build with engineers, so what ships matches the design. SKILLS Product design: Discovery, user research, usability testing, product strategy, information architecture, complex workflows, interaction design, conversational UX, AI product design (human in the loop), design systems, accessibility (WCAG), design QA Leadership: Leading project teams of 4 to 7, stakeholder alignment with founders and engineers, mentoring Tools: Figma, FigJam, Framer, Bubble, Claude Code, Codex, HTML, CSS, JavaScript, React, Git, Vercel Domains: AI, B2B SaaS, fintech, healthtech Languages: English (professional working proficiency) EXPERIENCE Product Designer and Design Engineer, Independent (freelance and contract), 2021 to 2025 Remote, for founders in India, the US and the EU * Led product design and front end delivery of products, MVPs, design systems and websites for founders, often running several engagements in parallel. Product Designer and No Code Developer, ACEAI CLUB (now Garage Labs), 2025 Mentorship marketplace for an AI focused alumni network * Took a nonfunctional Airtable prototype to a live, responsive platform: profiles, mentor discovery, scheduling and bookings. * Designed the activation journey (LinkedIn sign in, multi step profile) that turns a sign up into a bookable mentor, growing the platform from 0 to 2,300+ activated mentor profiles in 2 months (mentors who completed sign up and created a profile). Lead Product Designer, Cryptoventures, 2024 to 2025 Fintech: web3 investment platform for Indian retail investors (company since closed) * Redesigned a legacy, non responsive platform as a responsive product, replacing intrusive ads with native placements and an exit intent flow: the rates advertisers paid rose about 2.7x against the legacy site, and session time rose. In House Builder and Mentor, Era of No Code, 2024 to 2025 No code education company and builder network * Picked from the cohort by three mentors to build for founders: 19 projects (websites, landing pages, MVPs); steered first time founders to shadow test ideas before paying to build them. Later mentored cohort students. Founding Designer and Builder, Vinyasa, 2024 Healthtech: mental health product; shipped, and the original MVP was later sold to a US mental health startup * More than 75% of new users who started the onboarding questionnaire left before finishing it. From therapy transcripts and interviews I replaced it with a 2 to 3 minute voice check in and shipped the first functional MVP in 21 days. * Rejected an off platform audio conversion workaround that would have moved health data off the platform, and built the middleware on the platform myself (HIPAA aligned handling for US users). * Designed and built the practitioner platform (scheduling, notes, conflict detection, double booking prevention) and the Vinyasa Marketplace with licence verified provider onboarding. Designer, ConvertML, 2021 to 2022 B2B SaaS: GenAI customer analytics platform * Sole designer: brand identity (logo, guidelines) and the responsive marketing site, from scratch. PROJECTS * TmAi (2024, research led test product and interactive prototype): airline concierge subscription: research, business model, pricing tiers and a working conversational prototype. * Tanith Intelligence (2026, competition concept): AI native drone command system. Led product strategy, IA and UX for a team of four: 3rd of 100+ teams, 5th of 20 in the global jury round of FlytBase x Design Fusion. * Independent mobile apps (2026, testing with beta testers already): designed and built Start Anyway, unfold and Pour end to end. Also conversion audits and redesigns of two live SaaS products, Clacky and SeenPaid. ACHIEVEMENTS * Invited speaker at an official Framer workshop at IIT Delhi, teaching 93 attendees; hosted a community meetup at IIT Delhi on Figma semantics and design systems. Design Engineer resume (https://paatni.vercel.app/Abhishek-Paatni-Design-Engineer-Resume.pdf) SUMMARY Design Engineer with 4+ years designing and building products across 5 startups, from 0 to 1 and from 1 to 10: I grew ACEAI CLUB from 0 to 2,300+ activated mentor profiles in 2 months, and rebuilt an established platform with its developers at Cryptoventures. Often the founding or first designer, leading project teams of 4 to 7. I take a product from interactive prototype to production front end in React, TypeScript and Next.js: component architecture, design tokens, accessibility and automated tests, with AI assisted development (Claude Code, Codex) directed by written specs, keeping design intent intact from Figma to production. SKILLS Front end: HTML, CSS, JavaScript, TypeScript, React, Next.js, Vite, GSAP, responsive layout, accessibility (WCAG) Design to code: Component architecture, design tokens, Figma variables, interactive prototypes, design QA Integrations: OAuth, payments, Google Calendar sync, REST APIs, Node serverless functions, Gemini and Groq APIs, Flutter Quality and AI: Playwright, Vitest, Flutter test, GitHub Actions CI, Git, Vercel; Claude Code, Codex, specs and schema contracts Design: Figma, FigJam, Framer, Bubble, interaction design Languages: English (professional working proficiency) EXPERIENCE Product Designer and Design Engineer, Independent (freelance and contract), 2021 to 2025 Remote, for founders in India, the US and the EU * Designed and built products, MVPs, design systems and websites for founders in React, Framer and Bubble, often running several engagements in parallel. Product Designer and No Code Developer, ACEAI CLUB (now Garage Labs), 2025 Mentorship marketplace for an AI focused alumni network * Took a nonfunctional Airtable prototype to a live, responsive platform: profiles, mentor discovery, scheduling, bookings. * Implemented LinkedIn OAuth, payments and timezone aware two way Google Calendar sync, growing from 0 to 2,300+ activated mentor profiles in 2 months (mentors who completed sign up and created a profile). Lead Product Designer, Cryptoventures, 2024 to 2025 Fintech: web3 investment platform for Indian retail investors (company since closed) * Rebuilt a non responsive legacy platform as a responsive product, with an exit intent flow built with developers; native ad placements lifted the rates advertisers paid about 2.7x against the legacy site. In House Builder and Mentor, Era of No Code, 2024 to 2025 No code education company and builder network * Picked by three mentors to build for founders: 19 projects (websites, landing pages, MVPs) across design, development and product strategy. Later mentored cohort students. Founding Designer and Builder, Vinyasa, 2024 Healthtech: mental health product; shipped, and the original MVP was later sold to a US mental health startup * Shipped the first functional MVP in 21 days from early Figma frames: a voice check in that replaced an onboarding questionnaire more than 75% of new users left unfinished. * Built the on platform audio conversion middleware (WAV to MP3) myself instead of an off platform workaround, keeping health data on the platform (HIPAA aligned for US users). * Designed and built the practitioner platform: scheduling, notes, conflict detection and double booking prevention. Designer, ConvertML, 2021 to 2022 B2B SaaS: GenAI customer analytics platform * Sole designer: brand identity and the responsive marketing site, from scratch. PROJECTS * Start Anyway (independently built mobile app, Android beta, 2026; Flutter, Node): an AI planner that calls Gemini, falls back to Groq and validates every reply against a shared schema; about 1,250 automated tests in CI on GitHub Actions. * Vinyasa check in (2026 rebuild; Next.js, TypeScript PWA): privacy first voice analysis API, Web Audio soundscapes, Playwright tests that assert nothing is stored. * TmAi and Tanith sites (React, Vite, TypeScript, GSAP): design tokens, accessibility, build checks that fail when content is lost. Clacky and SeenPaid redesigns in plain HTML and CSS. ACHIEVEMENTS * Tanith Intelligence: led product strategy, IA and UX for a team of four: 3rd of 100+ teams, 5th of 20 in the global jury round of FlytBase x Design Fusion. * Invited speaker at an official Framer workshop at IIT Delhi, teaching 93 attendees. ## Full case studies, word for word from their own sites ### TmAi, Flight Concierge (source: https://tmai-casestudy.vercel.app/llms-full.txt) #### TmAi: Nobody makes money selling flights. A product design case study by Abhishek Paatni. Bangalore, 2024. Client: Rebecca Doshi. Role: Research, product design, prototype. Rebecca Doshi had an idea for airlines and one question: would it hold up? Four months later I came back with the numbers, the pricing and a working prototype. ##### 1. The brief ###### Her idea. My job was to test it. Rebecca Doshi asked one question. Could an airline sell a seat without paying somebody else to sell it? Airlines move billions of people a year and keep almost none of the money. Rebecca asked me to find the part of that business a designer could change, and to tell her honestly if there was nothing there. I came back with an answer neither of us expected, and it changed the whole project. > The problem was never the seat. It was the way the seat gets sold. ##### 2. The problem ###### Meet Arjun. I sat with eleven frequent flyers. Arjun is all of them at once. * Role: Sales lead * Flies: 4 to 5 times a month * Route: Bangalore to Delhi * Books: 2 days out, usually ####### Problem one: Seventeen tabs, and he still is not sure. The airline, two booking sites, then the airline again to see if direct is cheaper. By tab nine he has forgotten the first price. ####### Problem two: Then his accountant asks for the receipts. Every trip costs him twice: once to book, and again at quarter end, digging four months of confirmations out of his inbox. On a Sunday. He has come to hate Sundays. ####### Problem three: And his colleagues cannot pay for their own seats. Three people, one meeting. Arjun puts every seat on his card and chases the money, or they book apart and sit scattered down the aircraft. > Nobody wakes up wanting to compare fares. ##### 3. The solution ###### Seventeen taps become one. Six dropdowns, a results page, a fare table, a seat map, a payment sheet. Most people quit at the fare table, so I designed a conversation instead. ###### Edge cases: Ten ways a trip breaks, and what the agent does about each. A form waits for the traveller to find the problem. The agent finds it first and brings the fix. * 31%: Flights left late (India's four busiest airports, January to September 2024) * 3.67%: Flights cancelled (January 2024 fog, Indian domestic) * 7 days: To refund a card payment (The DGCA deadline for airlines) * 4.13%: Added to a dollar fare (Indian card markup, 3.5% plus GST) ####### Before booking * Visa restrictions. Every stop is checked against his passport using IATA Timatic. Routes he cannot take never reach the shortlist. * Policy violations. Fares over his company's cap are flagged before he picks one. His manager approves in the thread. * Seat inventory conflicts. Seat and fare are checked again at payment. If either has gone, he gets the closest match and approves any new price first. ####### At payment * Failed payments. The seat stays held at the same fare for thirty minutes while he tries another card or UPI. * Foreign currency payments. He sees and pays in rupees at the quoted rate, so the statement matches the screen. * Airline API failures. If the airline's system goes quiet, it says so. A charge with no ticket behind it is refunded automatically. ####### After booking * Delayed flights. It watches the incoming aircraft, so Arjun hears about a delay before the gate does, with a flight that still makes his meeting. * Cancelled flights. The news arrives with a seat already held on the next good flight. One tap keeps it. * Rebooking. The new ticket keeps his seat, his meal, and his colleagues seated together. * Refunds. It files the claim and tracks it in the thread until the money is back. Fare drops count too. Where these numbers come from: * Delays: DGCA data analysed by ThePrint. 109,607 of 354,775 departures at Delhi, Mumbai, Bengaluru and Hyderabad left late between January and September 2024. * Cancellations: DGCA monthly reports. 3.67 percent of scheduled domestic flights were cancelled in January 2024. Most other months sat between 0.6 and 1.9 percent. * Refunds: DGCA CAR Section 3, Series M, Part II, which gives airlines seven days to refund a ticket paid by card. * Currency: published card fees. Standard cards from India's largest banks add 3.5 percent to foreign currency spends, plus 18 percent GST on that markup. ##### 4. The research ###### An airline keeps three dollars and seventy cents of every hundred. IATA's 2025 forecast for every airline in the world: 979 billion dollars of revenue, and 36 billion of profit. * 3.7%: Industry net margin (2025 forecast, all carriers) * $7.20: Profit per passenger (Per flight flown) It is why the water bottle on board is 250ml and not 500ml. Every extra kilo in the air burns fuel that nobody paid for. It also means there is almost nothing left to cut inside the aircraft. If you want to change an airline's economics, the cabin is the wrong place to look. ###### The bill for selling the ticket is bigger than the profit. About half of all tickets sell through an agent or a booking system, and each one costs the airline around eight percent. * $39.16B: Paid to sell the seats (8% on $489.5B of indirect sales) * 61%: Already booked direct (T2RL, passengers boarded, 2022) Eight percent of 489.5 billion dollars of indirect sales is 39.16 billion a year, paid to companies that never flew anybody anywhere. That was the moment this stopped being a booking app and became a project about distribution. ###### One hundred dollars of ticket. * A hundred dollars of ticket revenue. All of it looks like the airline's money. * Most of it never was. Fuel, crew, leases, maintenance, airport charges. * Then the booking path takes its cut. A travel agent, a booking system, or both. * What survives is three dollars and seventy cents. * Sell it direct and that eight dollars stays. Same aircraft, same seat, same passenger. $3.70 becomes $11.70. Three times the margin, on a seat that did not change. Where these numbers come from: * IATA industry outlook 2025 for revenue, profit and net margin. * T2RL on distribution share, 2022, for the split between direct and indirect booking. * Everything else is modelled from public industry data and the assumptions written on this page, not from any single carrier. > The industry made 36 billion. It spent 39 billion getting the seats sold. ##### 5. The prototype ###### Sell it yourself, and make it worth buying. If the eight percent is the problem, stop paying it. People book wherever it is easiest, so direct had to become the easiest. Easier than seventeen tabs meant no form at all. You say where you are going and it books, because it already knows your seat, your meal, your card and who you travel with. Then I priced it as a membership, so the airline earns from a passenger even in months they never fly. ###### From one button to a booked flight. The agent lives inside the airline's own app. 01. Built into the airline's app. It sits where people already search and compare. 02. One button hands it over. Tap once, then talk to it like a person. 03. It does the searching. It asks what it needs and comes back with the three to five best options. 04. You choose, it books. Pick one. It books it and handles everything after. ###### It was all drawn in Figma first. The argument was about sequence rather than surface, so the flow came first. Every screen below came out of that file. At first the conversation carried everything as plain text, so a flight was a paragraph you had to read. It worked, and it was slow to scan. So every decision became a card the traveller could act on without leaving the thread. The plan screen got a clear recommendation, and the confirmation got the fare breakdown that later became the trust anchor in the live prototype. Screens: The old form, Handover, Choosing a plan, The conversation, Flights found, Confirmed, Adding people, Fare detail. ###### The final screens Six moments from the finished prototype, from the old form to a booked trip with the bill split. ####### 01 Before the handover * Six dropdowns before a single fare appears. * And four more screens behind the search button. * Every tap is counted. Most people give up around the fare table. * One button hands the whole trip to the agent. ####### 02 Choose a plan * The only choice before the chat begins. * Business leads, because it is built for people who fly every week. * Priced per year, so the airline earns in months nobody flies. * What each tier adds, in a few words, not a feature table. ####### 03 The chat opens * It types at reading pace, so replies land like a person's. * Passport, seat and card are already on file. Nothing to fill in. * The likely answers are one tap away. * Or say it in your own words. ####### 04 Flights found * It shows what it searched, so the shortlist can be trusted. * Three options instead of three hundred. * One full price per flight. No fare table to decode. * Picked for the trip, not just the route: landing before the first meeting. ####### 05 Split with colleagues * Colleagues are tagged with an @, the way people already chat. * Each colleague gets a payment request for their own seat. * Arjun fronts nothing for anyone. * Seats beside him are held at the same fare for thirty minutes. ####### 06 Booked * The boarding pass lands in the chat and the wallet. * Colleagues' payments are tracked in the same thread. * Distribution fee: zero. The airline keeps the eight percent. * Receipt emailed and filed for expenses. No Sunday lost to it. ###### Information architecture: One thread, and everything it needs on file. One button in the airline's app opens TmAi. Every decision happens in one thread, and the rest stays on file so Arjun never types it twice. * Airline app (Book a flight), then one button, opens TmAi: * Plan (Chosen once): Economy Plus, Business Class, First Class * The thread (Where everything happens): Flight cards, Fare breakdown, Colleagues and payment requests, Boarding pass, Delay, cancellation and refund alerts * Trips (To look things up): Upcoming, Past, Receipts, Quarterly expense report * Profile (On file): Passport, Seat and meal, Cards and UPI, Company travel policy, Colleagues Every extra tab is another place to look, which is the problem Arjun already has. So anything he has to decide arrives in the thread as a card he can act on. The other screens are for looking things up: a past trip, a receipt, the passport on file. None of them asks him to do anything. ###### User journey map: Arjun's trip, end to end. One Bangalore to Delhi trip, told twice. Grey is how it goes today. Blue is the same trip with TmAi. | Stage | Today | With TmAi | |, |, |, | | 01 Plan | A client meeting in Delhi on Thursday. An evening goes on travel. | He types one sentence. | | 02 Search | Seventeen tabs, and still not sure. | Three flights that land before the meeting. | | 03 Book | Every seat on his card, then he chases the money. | Colleagues tagged. Each pays for their own seat. | | 04 Fly | Digs the boarding pass out of his email. | The pass is already in the chat and the wallet. | | 05 Delay | Finds out at the gate and joins the queue. | Warned early, and already moved to a flight that makes the meeting. (31% of departures left late in 2024) | | 06 Expense | A Sunday spent digging out receipts. | Receipts filed as he goes. Sunday stays his. | ###### I also worked out the pricing. A yearly membership in three tiers, each priced against a job people already pay a travel agent to do. ####### Economy Plus, $100 a year This tier answers the seventeen tabs: one sentence books the trip. I priced it low on purpose. Its job is to move bookings onto the airline's own channel, where the eight percent stays with the airline. ####### Business Class, $700 a year Priced for Arjun. At his pace the fare is not what hurts, the disruptions are. So it rebooks before a delay, refunds a fare that drops, checks visas and turns a quarter of receipts into one report. ####### First Class, $2,000 a year Nothing is left to automate, so it sells attention: a person on the line in under a minute, any seat without a fee, a lounge on every segment. For people whose time costs more than the fare. > Sell it direct and that eight dollars stays. ##### 6. The outcome ###### The money is not in the seat. It is in the eight percent paid to sell the seat. That was the finding. Everything else came out of it. * Research: Market sizing, and the unit economics behind every number on this page. * Pricing: A three tier membership, priced against what a travel agent already charges. * Product: The conversation design, the plan gating, the structure and the edge cases. * Prototype: A working build the engineering team could put in front of real travellers. * Market read: Who would buy this, who never would, and the risks that sit outside design. ###### Three things this project taught me. Each one came out of the research, not the screens. ####### 01. Check the law before the layout. An airline steering travellers to its own seats is exactly what competition law in the US, the UK and Europe exists to stop. Now I find the legal edge before I draw. ####### 02. Map who has to say yes. It only worked if rival airlines shared one neutral agent, and each had a reason to defect. Now I map who has to agree, and why, first. ####### 03. Research should change a decision. One finding turned a booking app into a product about distribution. I judge research by what it changes. Competition law. In the US, the UK and across Europe, rules stop a large player from using its position to squeeze smaller ones. An airline that runs its own agent and quietly steers travellers to its own seats is doing exactly that. The moment TmAi favours the airline that owns it, it stops being a concierge and becomes a problem for a regulator. Game theory. To be useful it has to work across every airline, which means one integration they all agree to, serving the traveller and leaving the passenger data alone. Every major airline would have to sign up at once, and each has a strong reason to defect and keep the data. Nobody moves first, so nobody moves. Adoption has a ceiling too. My best guess is that five to ten percent of full service carriers ship something like this within three years, and under three percent of budget carriers try. A budget airline has no room for a paid agent. The maths does not work on a cheap fare. Neither wall is a design problem, which is exactly why a designer has to spot them early. > Research earns its place when it changes what gets built. ##### About the author Abhishek Paatni, Product Designer & Design Engineer. Portfolio: https://paatni.vercel.app/ Not affiliated with any airline. Last updated: 2026-11-08. Canonical page: https://tmai-casestudy.vercel.app/ ### Tanith Intelligence (source: https://tanith-intelligence.vercel.app/llms-full.txt) Tanith: AI native disaster response case study by Abhishek Paatni Source: https://tanith-intelligence.vercel.app/ This is the full text of the case study page, generated at build time. Tanith Overview #### Tanith: disaster response in 90 seconds, not 48 minutes. The whole case study in 30 seconds. An AI generated walkthrough of one 90 second response: detection, verification, coordinated dispatch, ground handoff. Scan in 10 seconds 48 MIN → 90 SEC Target response 3RD / 100+ Global qualification 47 / 50 Top jury score, tied 2nd of 20 9 DAYS Concept to shipped The problem Spotting a disaster is already fast. Coordinating the response is not. The brief named a 48 minute gap between detection and a coordinated response on the ground, and asked for 90 seconds. At Mati, Greece, that gap cost 104 lives. The reframe Most teams redesigned the dashboard. I redesigned what the operator is for. AI is the agent. The human is the judge. What I built Tanith, an AI native command system for critical airspace. The AI detects, verifies and lines up the whole response. The human approves anything that can't be undone. Two states: a quiet Idle overwatch, and a Focused alert view that shrinks the screen to one decision. Every AI call, operator action and failsafe goes into a legal audit log. How the 90 seconds work * T+0s A sensor trips. Three AI layers clear the noise. The quickest drone launches, not the closest. * T+30s Drone in transit. Four feeds spin up: night vision, thermal, 4K, and a map with the trajectory. * T+60s The AI checks the feeds against past wildfires and live weather. Confidence climbs. * T+90s Wildfire confirmed at 94%. The operator proceeds or aborts, or a 25 second failsafe proceeds and logs it. Ground teams are already moving. Four decisions I owned, and what each one cost * 01 A human judge instead of full autonomy. Cost: Seconds of latency on every irreversible action. * 02 A new company instead of the rebrand the brief asked for. Cost: Jury points. I'd make the call again. * 03 Dead Man's Switch: four escalation layers, not one. Cost: Complexity. * 04 Send the quickest drone, not the closest. Cost: Explainability. What I owned Brand, product strategy, information architecture, UX, the 90 second decision chain, and the live Framer site, which I wrote, designed, built and shipped alone. I led a team of four. The read that mattered most "You people have gone way outside the problem statement and truly tried to work out a solution that can actually be called an AI native solution with human in the loop. Nobody in this designothon or designothon before this ever worked like that." Prathmesh, Lead Product Designer, FlytBase ##### The brand and the product at a glance. Tap any image to expand it. Conference signage, built for the screens where govtech contracts get decided. The monogram on hardware. Trust starts at the airframe. Commanding Autonomy in Critical Airspace. Built for emergency operations, surveillance and critical missions. Three messages, three audiences, one identity. The Overwatch state. Drones at rest, heat halos showing regional risk. Silence is the system's default, and its safety signal. A sensor trips at 3:47 AM. The drone is in the air before the operator's screen finishes flashing. Four feeds, one conclusion. 94% certainty and a 25 second countdown. The Dead Man's Switch at work. Three units, three ETAs, one bulk dispatch. The system assembles the response. The operator approves it. ##### Detailed case study begins here July 23, 2018 · Mati, Greece ###### The fire reached homes before help did. 104 people died, most of them in the first 90 minutes. Tanith is the command system we designed to close that gap. The brief put the average time from detection to a coordinated response at 48 minutes. We designed for 90 seconds, in 9 days, as a team of four that I led. 9 Days 3RD / 100+ Qualification 5TH / 20 Final Jury Use the menu at the top right to jump to any chapter, or follow Up next at the end of each one. Watch the AI concept video ##### We didn't redesign the dashboard. We redesigned what the operator is for. Every team got the same brief: take 48 minutes down to 90 seconds. Most of them went to work on the dashboard. We didn't. Legacy command software treats the human as the pilot. They navigate, read the data and decide what comes next. That made sense in 2015. Today, drones can detect, launch and navigate on their own. At 3:47 AM the human isn't a pilot. They are the bottleneck. So we flipped the role. * AI is the agent. It detects, launches, navigates and gathers the evidence. Watching many feeds at once and finding a path in milliseconds are things people are bad at, especially under pressure. * The human is the judge. They check what the AI found, decide whether the confidence is high enough, and approve anything that can't be undone. They carry the legal accountability, and they can overrule the system. Every screen in Tanith follows from that split. The human brain is a decision engine, not a search engine. Pair operations: a senior commander and a supporting officer, the wingman setup our research found in emergency rooms, air traffic control and military command. One tired operator at 3:47 AM is a failure mode, so the system assumes two. ###### The harder reframe: a new company, not a rebrand. The brief asked for a FlytBase rebrand. We delivered a new enterprise. A company filed under drone fleet management software doesn't make the bidder lists for defense, govtech and emergency services contracts. Those lists are gated by category, relationships and security posture (full reasoning ). So we built a company that could walk into those rooms, and named it Tanith, after the ancient goddess of safe passage. The jury docked points for that decision. The FlytBase lead designer thanked us for it. ##### What I owned. Who I built it with. I led a team of four. We had never worked together before day one. Saniya and Lakshmi were design students. Vanshika was a UI/UX designer about a year into her career. That team shipped Tanith in nine days. * Brand strategy and positioning. The name, the new enterprise reframe, the archetype, the tone of voice. * Product strategy and IA. The Idle and Alert states, the human as judge model, the Dead Man's Switch, the edge case map. * UX flows and decision logic. The 90 second chain, the confidence thresholds from research (Stack 5 ), and quickest drone dispatch. * The Framer website. Written, designed, built and shipped alone, in parallel with everything above. * AI image generation. Atmosphere, pitch material and the future vision shots. * Team leadership. Strategy, settling disagreements, holding scope, and delivering against a nine day clock. Day 2, 5:07 PM. Presenting the gameplan to the team, with the name shortlist on screen (Tanith, Hedron, ORDIN). The rest of the team built against this plan for seven days. ##### What we found. Not what we did. ###### The problem statement. A single operator must reliably detect, verify, and coordinate life saving drone actions across multiple zones under extreme time pressure and uncertainty, while the system earns and maintains trust in every second. Nine days left no room for research theater. We built five evidence stacks, each backed by named sources. All five pointed at the same root cause. ###### Stack 1: the competition. Every player solves part of the response loop. Nobody owns the whole chain, from detection to a coordinated ground response with a human judging. The gap is the chain itself. Who we looked at * DroneSense. Drone fleet management, no decision intelligence. * Skydio. Defense grade hardware and autonomy. * Paladin Drones. Drone first responder dispatch for public safety. * Auterion. An open source autopilot operating system. * Palantir Gotham. The benchmark for enterprise data fusion. ###### Stack 2: how high stakes interfaces work. We studied 15 interface patterns from air traffic control, military command, nuclear control rooms, surgical dashboards and 911 dispatch. They agree on five rules: silence means safe, confidence shows as a range, irreversible actions wait for a held confirmation, alerts sort by severity instead of source, and the operator's mental load is the constraint that matters. Four of the fifteen patterns * STCA (Short Term Conflict Alert), air traffic control. Graduated alerts instead of on and off. * MIL STD 2525, NATO command and control. A "halo of ambiguity" around uncertain hostile contacts. * Westinghouse Ovation "Shelve", nuclear control. Alarm suppression with a full audit trail. * RapidSOS, 911 dispatch. Incident data passed to first responders with its confidence attached. * Eleven more, each documented. ###### Stack 3: how real drone operators read screens. Experienced operators look at the threat feed first, telemetry second and mission status third. Novices do the reverse. Interfaces built on the expert order cut decision time by about 40% in published trials, so we built on it. Sources A Springer eye tracking study comparing novice and experienced UAV pilots. Advexure's field guide to public safety drone operations. An arXiv field trial paper on firefighting with autonomous drones. ###### Stack 4: how these systems fail. The biggest failure isn't technical. It's the moment the operator freezes, asked to judge something they aren't ready to judge. In air traffic control simulator studies, pilots routinely ignore low priority AI flags. In a fire, that is lethal. The Dead Man's Switch exists because of this finding. Sources Ten documented failure scenarios, drawn from CAA UK drone safety reports, MDPI studies on vision AI fire detection, NCBI work on GPS spoofing, anvil.so on human error in drone operations, and anikedia research on drone UI overload. ###### Stack 5: what makes people trust AI. Operators trust AI when it shows confidence as a range, not a single number. The thresholds we took from the literature: below 65%, the AI asks and the operator decides. From 65% to 84%, the AI suggests and the operator approves. At 85% and above, the AI acts and the operator keeps the override. AI SUGGESTS. Operator approves or rejects. PROCEED Drag the slider. The system only acts on its own when confidence clears 85%. These bands decide whether Proceed is dimmed, highlighted or counting down, and how long each decision holds. Sources CDC crisis communication research, NASA studies on handing off control, EU civil protection coordination protocols, and NTSB findings on collision avoidance. ###### The root cause. The operator at 3:47 AM isn't failing because the screens are wrong. They're failing because the role is wrong. They're asked to be pilot, analyst, dispatcher, witness and judge at once. Nobody does all five well in the first minute of an emergency. So the redesign starts with the role, not the screen. Day 3, 6:38 PM. Lakshmi walking the team through the competitor scan, with DroneSense, Paladin Drones and Auterion on screen. Research was something we argued over together, not a solo doc. ##### Three options. One choice. The tradeoff named. Every real design decision gives something up. These are the four that shaped Tanith, and what each one cost. ###### Who acts: the AI or the human? Option A Full autonomy The AI detects, decides and acts. Humans review after the fact. Roughly where Anduril Lattice is heading. Option B Human driven, AI as support The operator does everything and the AI surfaces information. Where most agencies are today, and roughly what Palantir Gotham serves. Option C The AI investigates and lines up the response. A human approves anything irreversible. The Tanith path. Everything else keeps running on its own, under human oversight. Why not A Liability can't legally or morally sit with an AI. Insurance, evidence chains and post incident audits all need a named human. A regulator can subpoena the operator. They cannot subpoena the model. Why not B This setup is what causes the 48 minute delay the brief asked us to fix. Human driven coordination doesn't scale to supervising autonomous systems. Why C, and what it costs Latency on irreversible actions. Holds are measured in seconds: about 5 for a small call, 25 for a fire dispatch, longer for bigger ones. That buys a human who answers for every irreversible step. We revisit when the law lets an AI be the accountable party. That is not 2026. ###### A new company, not a rebrand. The brief asked for a FlytBase rebrand. We weighed three ways to answer it. Option A FlytBase rebrand Same company, new identity, Tanith as a module. Cheapest, and closest to the brief. Option B FlytBase sub brand A product brand inside FlytBase, like "FlytBase Tanith". The middle path. Option C An entirely new enterprise New company, new positioning, new go to market. FlytBase's drone automation stays underneath as infrastructure. Why not A Drone fleet software doesn't appear on the bidder lists for defense, govtech or emergency services contracts. Category is a procurement gate, and a new storefront doesn't change the category in a contracting officer's eyes. Why not B A sub brand inherits the parent's category. To a defense buyer, "FlytBase Tanith" still reads as drone software. Why C, and what it costs Capital. A new enterprise costs far more than a rebrand: brand, positioning, go to market, regulatory posture and security clearances, all from zero. The jury docked us for exactly this, calling it harder to deploy. We held the position. The defense, in one paragraph. The cheap option fails earlier. A company in the wrong category loses every bid it enters. One in the right category competes from day one. The jury's worry about deployability doesn't survive that comparison. The FlytBase lead designer, who works inside that structure every day, saw it. We revisit when drone software vendors can credibly sell AI command intelligence with human oversight directly. Commanding Autonomy in Critical Airspace. The booth a defense procurement officer walks past at a govtech conference. Drone management software doesn't get this booth. ###### The Dead Man's Switch: when the operator doesn't respond. Operators go missing for reasons emergencies don't wait for: asleep at 3 AM, in the bathroom, on another call, in shock (Stack 4 ). The system needs a fallback that is neither full autonomy (rejected above ) nor a deadlock. Option A Earthquake chair The chair shakes and sounds an alarm when a critical alert goes unanswered. Option B Smartwatch escalation The alert moves to the operator's wrist when they're away from the desk, with haptics, a flash and sound coded by severity. Option C Peer escalation A free commander in another zone gets control and joins the comms loop. Option D Calibrated auto proceed No response within a window sized to the decision? The AI takes the recommended action and logs the failsafe. Fatal actions only auto proceed above 85% confidence (Stack 5 ). Why not just one Each one fixes one failure and leaves the rest open. The chair fails if the operator is in shock at the desk. The watch fails with no signal. Peer escalation fails when nobody is free. What we chose All four, stacked. D is the floor that always happens. A, B and C sit above it: the chair, then the watch, then a peer, and only if all of them fail, the auto proceed. What it costs Complexity. Four layers are more to design, test and break. Picking one would mean accepting a known failure as the default. In this product, we wouldn't. Layer 2 of the Dead Man's Switch: the wrist alert when the operator is away from the desk. ###### Send the quickest drone, not the closest. The usual answer is the closest drone. Tanith sends the one that will arrive first. They are often different. A drone already in the air can turn and get there faster than a docked one that is nearer but has to undock, climb and speed up. Against a 90 second target, that gap can be 15 to 40 seconds. In a wildfire, it's the difference between catching the fire 50 m from homes and 200 m from them. The math is the AI's job. The operator sees which drone is going, its ETA, and why, if they ask. The eight variables, checked for every drone in under a second * Altitude and heading. * Vertical and horizontal speed. * Distance from dock, for docked drones. * Terrain between each drone and the site. * Weather between each drone and the site. * Fuel remaining. * Sensor payload match for the type of anomaly. No human can weigh eight variables across twelve drones in under a second. The AI can. That is what AI native dispatch means. What it costs: explainability When the AI skips the obvious drone, the operator may not see why. So the screen shows both ETAs: the closest drone's and the chosen one's. What the AI weighs for each drone: altitude, vertical and horizontal speed, latency, distance from dock. Hover any drone to see them. ##### What we built. The system first, then the screens. Here is how Tanith is organized, then one incident followed through it. ###### Two states, and a funnel between them. In Idle, the map is calm on purpose: drones docked or on patrol, risk halos over regions, live weather. Calm is the safety signal, so when something changes, it means something. In Alert, the map collapses into the Focused screen: four feeds, the AI's assessment, and one decision. Between them sits the Funnel of Truth. Three AI layers filter false positives, match the event against past anomalies, and check confidence against the trust thresholds. Most of what reaches a sensor never reaches the human. The human is too expensive to spend on noise. Underneath it all runs a legal audit layer that timestamps every telemetry frame, AI recommendation, operator action and auto proceed. A system this autonomous can't ship without one. ###### The information architecture. Idle, then Alert in two phases: assess, then respond. The screens inside each, and two layers that never switch off. Tanith command console * Idle state ####### Overwatch * Overwatch map: Drones docked or on patrol, risk halos by region, live weather. * Drone detail, on hover: Altitude, speed, distance from dock, battery, live feed. * Sector status: Monitored, unmonitored, or covered by a patrol drone. * Alert state · Assess ####### Focused screen * Four live feeds: Night vision, thermal, 4K camera, map with trajectory. * AI verdict: Anomaly type, with confidence shown as a range. * Situational report, on demand: Wind, temperature, humidity, visibility, pressure. * Decision bar: Proceed, Abort, and the hold timer. * Alert state · Respond ####### Dispatch and coordination * Deployment panel: AI strategy, suggested units with ETAs, Dispatch All. * Spread projection: Predicted path, zones and homes at risk. * Ground team channel: Routes, ETAs, comms per team or all teams. * Active response: Drone feeds to crews, evacuation guidance by light and speaker. * Idle → Assess: An event clears all three AI filters. * Assess → Respond: Proceed, by the operator or the failsafe. * Assess → Idle: Abort. The drone flies back to its dock. * Respond → Idle: The incident closes and is archived. Always on Legal audit layer. Every AI call, operator action and failsafe, timestamped and archived. Dead Man's Switch. Chair, then watch, then a peer, then a calibrated auto proceed (why ). The full flow map from the sprint. Passive state, focused screen, situational report, deployment, path projection, ground coordination, active response. Abort sends the drone back to dock. Legal audits run under everything. ###### The 90 second chain. Four beats. Each one is a screen. T+0s Sensor trip. What happens Edge sensor 42 spots a thermal anomaly in Sector 04. The event clears the Funnel of Truth, and the AI launches the quickest drone in the region, not the closest. Idle flips to Alert and the sector turns red. DRONE 42A is already chosen. The operator never has to pick a drone. T+30s Drone in transit. Four feeds spin up. What happens Drone 42A is airborne. The Focused screen opens with four feeds: night vision, thermal, 4K camera, and a map with the trajectory. Nobody asks the operator anything yet. The system is bringing them the evidence. Feeds live as Drone 42A reaches the north west perimeter at 34.02, minus 118.45. The map quadrant shows trajectory and ETA. T+60s AI assesses. Confidence climbs. Thermal and video confirm the heat. The AI checks it against past wildfires and live weather (wind 13.4 m/s). Confidence passes 65%, then 75%, then 85%, where Proceed unlocks. T+90s Wildfire confirmed at 94%. Proceed lights up and a 25 second hold starts. The operator can proceed early, abort, or let the timer run out. ###### The held moment. Every irreversible action in Tanith passes through this screen. What each part of the screen does * "Wildfire Detected" is the AI's conclusion, not its inputs. * "94% Certainty" shows as a number and as a range bar, because operators trust ranges more than absolutes. * "PROCEED" is highlighted because confidence is above 85%. * "The Mission will automatically proceed in 25 sec" is the hold, sized to this decision. * "ABORT MISSION" is always available and never times out. * The feeds stay visible behind it, so the operator can check the evidence in the seconds they have. The hold scales with the stakes. 25 seconds lets an experienced operator glance at four feeds and confirm. Confirming a noise filter might hold for 5 seconds. Ordering an evacuation might hold for 60 or more. Need more context? The AI situational report opens next to the countdown without pausing it. The high risk variant. Wind 45.2 km/h, air 48°C, humidity 12%, visibility 1.5 km, pressure 1013 hPa, tagged CRITICAL. The decision comes first. The evidence is one tap away. If nobody acts in 25 seconds, Tanith proceeds, logs the failsafe and records every input behind it. That is the floor of the Dead Man's Switch . ###### Handing off to the ground. Once the operator (or the failsafe) proceeds, Tanith stops assessing and starts responding. What each part of the panel does * "PROCEED TO DISPATCH" keeps the same action verb across the whole chain. * "Incident ID: WF 0428" gives every incident a reference for the audit trail. * "AI Strategy: Tier 2 Fire Suppression + Zone 4 Evacuation" states the plan in plain language before any unit moves. * Three ready units. Fire Suppression, Engine Co. 12, Station 04, ETA 04:12. Civil Defense Unit 09, ETA 06:45, for The Oaks residential block (14 houses). EMS Unit 05, ETA 05:20, staging for heat distress. * "Dispatch All Teams" is one bulk action, because operators do best with two or three strong choices, not long lists. Three units, three roles, three ETAs, one Dispatch All. The AI assembled the response. The operator approves it. Crews get everything on their tablets: GPS, the projected spread, approach routes, wind, evacuation status and comms channels. Nobody arrives blind. The receiving end. The AI saves crews the 8 minutes they would spend radioing for basics. The operator's 90 second call becomes the crew's 8 minute head start. ###### Plugging into real sensors and drones. Designed against real hardware and open standards. Not wired to live hardware in nine days. * In. Tower thermal cameras that detect on the device, satellite hotspot feeds (NOAA, NASA FIRMS), ground smoke and weather sensors. Each becomes one standard event, so a new sensor needs an adapter, not a redesign. * Fly. Dock based drones through FlytBase's existing fleet layer. Tanith picks the drone and the mission, and FlytBase flies it. Other fleets connect over MAVLink. * Out. Dispatch goes into the agency's existing dispatch software as standard CAP alerts, and to crew tablets. Tanith feeds the dispatch center. It doesn't replace it. An agency starts with the cameras and satellite feeds it already has, then adds docks where risk runs highest. Partial coverage still works. It just verifies slower. ###### When things fail. A system this autonomous is judged by how it fails. The rule: degrade toward a human, never toward silence. Every failure is shown to the operator, has a fallback, and is logged. When a sensor fails * Sensor goes dark. The sector is marked unmonitored and the nearest patrol drone covers the gap. It shows as a status change, not an alarm. * False positive. The funnel filters most. If a drone disproves one, the alert closes with a logged reason. * Sources disagree. Nothing auto proceeds until a drone has looked. The drone goes anyway, because a look is cheap. When a drone fails * Link lost. The drone returns on its last safe path, and the next quickest drone launches at once. * Low battery or high wind. A drone that can't finish safely is never picked. If none can fly, cameras and satellite verify, and the operator is told why. * GPS spoofed. Position is cross checked on board. On a mismatch the drone returns and a security incident is logged. When the AI is unsure * Below the 85% threshold. Proceed isn't highlighted and fatal actions never auto proceed (Dead Man's Switch ). The system asks for a second look and the operator keeps deciding. When the human fails * No response. The Dead Man's Switch takes over. * Wrong call. An abort is never final: the drone stays on station and alerts a peer if confidence climbs. A wrong proceed can be recalled in one action. When the system fails * Console or network down. A standby console takes over, drones in the air finish on board, and alerts reroute to peers and smartwatches. It degrades to manual. It never goes quiet. That is what 90 seconds buys: a coordinated response already moving while legacy systems are still gathering their first sitrep. ##### User journey: one fire, one night shift. The duty commander, 3:47 AM. A wildfire starts in Sector 04. What they do, what Tanith does, and how it feels, next to how the same night goes today. Before T+0 ###### Overwatch Operator does: Watches a quiet map. Nothing moves unless it matters. Tanith does: Filters sensor noise through three AI layers and stays silent. Without Tanith: Alarms all night. Real alerts get lost in the noise. Operator feels: Calm T+0s ###### Alert Operator does: Sees the screen flip to Alert. One sector turns red. Tanith does: The event cleared all three filters. The quickest drone is already flying. Without Tanith: 6 min before a human even hears about it. Operator feels: Startled T+30s to T+60s ###### Verify Operator does: Watches four live feeds: night vision, thermal, 4K, map. Tanith does: Checks the feeds against past wildfires and live weather. Confidence climbs to 85%. Without Tanith: 14 min of calls and cross checks. Operator feels: Unsure, then steadier T+90s ###### Decide Operator does: Proceed or Abort. One decision on screen. Tanith does: 94% confidence. Proceed highlighted. A 25 second hold, then the failsafe. Without Tanith: 5 min to decide, often alone, often frozen. Operator feels: Peak pressure T+90s onward ###### Dispatch Operator does: One action: Dispatch All. Tanith does: Built the plan while the drone flew. 3 units, ETAs 04:12, 05:20, 06:45. Without Tanith: 18 min coordinating units by radio, then 5 min to dispatch. Operator feels: Back in control After dispatch ###### Handoff Operator does: Talks to crews, monitors, closes the incident. Tanith does: Loads crew tablets with GPS, spread map and routes. Logs every action for audit. Without Tanith: Crews arrive blind and spend 8 min radioing for basics. Operator feels: Calm again "Without Tanith" splits the brief's 48 minutes by our model (see the projections ). Tanith's times are design targets, not measurements. The stress peak doesn't disappear. It shrinks, from 48 minutes of chasing information to one 25 second decision with the evidence on screen. ##### Projected, not measured. Tanith was never deployed, so nothing here is measured. These are projections and test targets, labeled as such. ~3 Projected minutes, expected case minus 93% Projected vs the 48 min baseline 8 Projected minutes, worst case ###### Projected response time. The brief's 48 minutes, split into stages, then run through Tanith. ``` STAGE LEGACY TANITH (brief) (projected) Detection reaches a human 06:00 00:10 Verification 14:00 02:30 Coordination 18:00 in parallel Decision 05:00 00:25 Dispatch 05:00 00:10 TOTAL 48:00 03:15 ``` The saving doesn't come from one faster step. Legacy response verifies, then coordinates, then decides. Tanith does all three at once: the drone launches at the alert, and the dispatch plan is built while it flies. The assumptions * The 48 minutes is the brief's figure. The stage split is our model. * Both columns stop at dispatch. Travel time for ground teams isn't counted. * The nearest docked drone is about 2 km out, cruising at about 60 km/h. * The operator uses the full 25 second hold. ###### Operators under pressure. Not yet tested with real operators (what's next ). What exists is a design aimed at how pressure breaks people, and a test to prove it. * Tunnel vision. One decision on screen at a time. * Freezing (Stack 4 ). A held deadline, then the Dead Man's Switch. * Alarm fatigue. Noise is filtered out. Idle stays silent. * Choice overload. Three ranked units, one Dispatch All. * Distrust of the AI. Confidence as a range, and the closest drone's ETA beside the chosen one. ###### The pressure test. Two rounds of six dispatch operators on a simulated night shift, with four scenarios: a real fire, a false alarm, a lost drone, and a second alert in the middle of a decision. ``` MEASURE PASS MARK Time to decision after alert ≤ 25 s Correct calls (proceed / abort) ≥ 95% Missed critical alerts 0 SUS usability score ≥ 80 (industry avg 68) Follows the AI on low confidence < 10% ``` The last row matters most. Operators should trust the AI when it is sure and question it when it isn't. ###### From projection to measurement. A 30 day shadow pilot with one agency. Tanith runs silently beside the real dispatch center, and the agency's own response times are the control. ##### A brand that opens doors before the product does. Because Tanith became a new company (why ), the brand had to do work the product couldn't yet. In three seconds, a defense procurement officer had to think "serious vendor", not "student competition project". Six choices got us there. None of them were about taste. Visual direction Typography Color The monogram ###### The name. Tanith, the Phoenician goddess of safe passage, chief deity of Carthage, invoked when caravans crossed dangerous ground. Invented names read like agency output. A name with two thousand years of meaning reads as inherited authority. It ends on a hard consonant that lands when spoken, and it doesn't clash with any aerospace or defense trademark in the markets that matter. The two finalists that lost * Hedron. Invented and easy to read, but it suggested geometry more than gravity. A software name, not a defense one. * ORDIN. Too close to "ordinance" without a clear meaning of its own. It risked sounding like a procurement category instead of a company. Tanith won because it already meant something. The others had to invent it. ###### The monogram. Three stacked tiers: a chevron on top, two arrows in the middle, a diamond at the base. Read top to bottom, it is the AI watching from the sky, the operator's decisions in transit, and the ground the system protects. It is also a drone's view, looking straight down. The full exploration. Ten variants, reference scans of Spectra, Airspace and other defense identities, sketches and refinements. The stacked chevron and diamond locked on iteration 8. Tanith Full lockup. Custom letterforms paired with the monogram. ###### Color. Indigo, #4927B0. Blue is what every defense and enterprise brand already uses. Red is taken inside the product, where it means danger. Indigo is neither. Why not blue, why not red * Blue is the category default. Walk a govtech conference and twenty booths in a row are some shade of it. Standing apart cost nothing, so we took it. * Red already marks critical alerts, aborts and hazard zones in the UI. A red brand on a red alert screen would make an operator wonder whether something is the logo or on fire. * Indigo is cool and authoritative without being blue. It nods to Tyrian purple, the imperial color of the Phoenician world, it holds up on the dark UI, and it keeps the cool trust register defense buyers expect. ###### Typography. Three typefaces, one job each. * Space Mono for display, 48px and up. Monospace reads like instruments: air traffic control, telex, terminal output. The cockpit, not the marketing deck. * Inter for UI, under 48px. Legible at small sizes on screens, for an operator reading at 3 AM in a dark room, sometimes through tears. * Source Code Pro for data and telemetry. Zero never looks like O, and one never looks like l. That matters when you read sensor IDs, GPS coordinates and timestamps. ###### Visual direction. The brand had to say three things: speed, autonomy, authority. Everything else got cut. The rules we locked * Graphics. Simple geometry only: map lines, circles, arcs, cones, straight lines. Instruments, not ornament. * Illustration. Clean vector diagrams, top down or isometric, with no characters. Only for explainers, decks and onboarding. * Motion. Straight or curved paths only. No bounce, no elastic. Things slow down like heavy objects, not friendly mascots. * Materials. Matte metal, the language of aircraft and critical infrastructure. Subtle grain, so it never looks sterile. Directional depth to guide attention without panic. One soft glow, kept only for the AI's active state. ###### Touchpoints. Software is one surface. Hardware, wearables, print and signage are the rest. Drone airframe. Monogram and accent stripes. A brand built for hardware. Operator station boot screen. "Command. Decide. Act." Conference LED. "Autonomous Command Systems. When Every Second Matters." Govtech expo booth. "Commanding Autonomy in Critical Airspace." Three poster gallery wall. Three messages, three audiences, one identity. Public billboard at night. The monogram and the category, "Disaster Response System". Operator wearable. Layer 2 of the Dead Man's Switch. Personnel ID. The brand on a body, not just a screen. Operator uniform. The brand goes where the crew goes. Field cap. Same logic. The brand follows the person. Headquarters signage. The brand at home. Building exterior. The brand at city scale. Day 3, 7:21 PM. Lakshmi building the final monogram in Illustrator, vector by vector, against the visual rules above. ###### Brand archetype Hero, with a Sage influence. A hero's posture, the protector who acts when others can't. A sage's voice: calm, evidence led, never shouting. Defense buyers and emergency service directors take that counterpart seriously. ##### The website. Written, designed, built and shipped solo. ###### tanith.framer.website When Every Second Counts, Tanith Responds. Visit the live site The live site at tanith.framer.website was the day nine deliverable. I built it alone in Framer: copy, design, layout and deployment. Why solo: a brand voice needs one editor. Built by committee in nine days, the voice and rhythm fracture. Tanith sounds calm, evidence led, and never longer than it needs to be. That took one person holding the pen. Why Framer: design fidelity and deployment speed in one tool, against a hard submission deadline. Its job: read like a vendor a defense contracting officer could route an RFP through. A company, not a portfolio piece. The blocks on the page, and what each one is for * Hero. "When Every Second Counts, Tanith Responds." The category claim and the difference, in one sentence. * About Us. "Born From Disaster. Built To Protect." Why this company exists. * Our Aim. Purpose, mission and vision, the institutional voice govtech buyers expect from a vendor. HR readers can skip it. Procurement readers can't. * How It Works. Detect, Confirm, Act. The 90 second chain in three beats for someone scanning. * Scenarios. "Operator Centered, Not AI Obsessed." Four product UI tiles, each with a one line principle: feed and context for fast decisions, a confidence based call to action, the overwatch state, live comms. * Why Us. Six principles, one sentence each: Faster Than Human Reflexes, Clarity Over Complexity, Confidence Transparency, Built for Real Environments, Human In The Loop Control, From Detection to Delivery. * Testimonials. Four positioning quotes from four personas in four regions: California fire ops, a Berlin hospital ER, an emergency drone unit operator, Japan flood response. Each one speaks to a buyer segment. * Solution video. The same AI generated walkthrough that opens this case study. * Onboarding. The 30 day rollout: site audit, operator training, live deployment. Procurement needs to know what buying this looks like. * FAQ. The six questions a procurement officer or head of operations asks before signing anything. Every block answers a buyer's question. Process pages, designer credits, methodology and color palettes stayed off the site. They live here instead. The case study is for hiring managers. The website is for buyers. ##### What the jury said. What it cost. What I'd do again. ###### The numbers. ``` Qualification stage: 3rd of 100+ teams globally. Final jury (top 20): 5th of 20. Final score: 82 / 100 (Rajat 35 / Nisha 47) Winning team (KRANTI): 93 / 100 (Rajat 44 / Nisha 49) ``` The final scoring sheet from the jury: top 20 teams, two judges, out of 100. Certificate of Merit, Design Fusion Season 3, signed by Vannsh Agrawal, founder and CEO of Designare, for the top 20 finish. ###### On the record. "The solution, UI, flows, IA and thinking behind the brand and product looks grounded in reality. I feel like this is something that can be shipped to developers for development and then shipped to the world, and I loved the fact that you people did not try to make it pixel perfect and gimmicky." Rajat Patel, Jury Member, Design Fusion Season 3 "I never expected any team from the designothon to ever come up with a solution that can be pitched in real life to investors by navigating real world regulations, laws and constraints while using the current UI, flows and design system of FlytBase. You people have gone way outside the problem statement and truly tried to work out a solution that can actually be called an AI native solution with human in the loop. Nobody in this designothon or designothon before this ever worked like that." Prathmesh, Lead Product Designer, FlytBase(quote used with permission) Nisha Kumari's read is about the pitch, so it sits with the pitch below. ###### The honest read on 5th place. Nisha scored us 47 of 50, tied for her second highest score in the top 20. Only the winners, KRANTI, scored higher with her (49). On storytelling and clarity, we were effectively in the top two. Rajat scored us 35 of 50, below average for the top group. He told us his lens included whether the solution could realistically deploy, and the new company decision (why we made it ) added cost to deploying it. Would I make the same call again? Yes. We knew the decision would cost us and held it anyway. Drone management software doesn't make the procurement lists this product needs, and a company in the right category is cheaper than one that loses every bid. So 5th place is the price of a position the jury didn't fully accept and the FlytBase lead designer did. Operators, engineers and product leaders judge substance. Competition juries judge fit to the brief. I think that trade favors the job I'm after. ##### Two minutes. Three disasters. One solution. The final jury round of Design Fusion Season 3 was a live pitch of exactly two minutes. Run over and you lose points. Run under and you lose conviction. Nine days of work had to fit in less time than most TikToks. ###### How it opened: three real disasters, one beat each. Mati, Greece, 2018. 104 dead, most in the first 90 minutes. Officials were later convicted for failures in the response. Pedrógão Grande, Portugal, 2017. 66 dead on a single road, trapped as the fire crossed their evacuation route. It became known as "The Road of Death." Black Summer, Australia, 2019 to 2020. 33 dead, 3 billion animals killed, 24 million hectares burned. The largest fire event in Australia's recorded history. Three countries, one pattern: the response was slower than the fire. ###### How it built. Then the chain: detection, AI assessment, the operator's call, coordinated dispatch, crews moving. The same 90 seconds from the solution , told as a wildfire the jury could feel happening. Then the brand, the deck and the close. Two minutes flat, and no questions needed. ###### What the jury said. "The storytelling of the pitch to explain a scenario that not only tells us the impact of the work you did but also shows how you were actually able to solve the 45 minute rescue delay to 90 second response, alert and rescue operation, keeping the pitch under 2 minutes but exact 2 minutes was amazing. Nobody has been able to do that and the fact that I do not have to ask any question to understand your solution is also wonderful." Nisha Kumari, Jury Member, Design Fusion Season 3 This quote means more to me than the placement. Squeezing a whole system into 120 seconds without losing the story or the rigor is hard. Nisha is saying we kept both, clearly enough that nobody on the panel needed to ask a question. ##### The dent Tanith was designed to make. Third party data only. Every figure is sourced, and where there is an estimate, the method is named. ###### Detection is fast. Response isn't. NOAA's Next Generation Fire System, live since February 2025 and used by 90% of National Weather Service forecast offices, can spot a new wildfire from satellite imagery and alert authorities in as little as one minute [Source: NOAA, May 2025]. In Oklahoma's 2025 Great Plains outbreak, up to 80% of satellite detected fires reached authorities before any 911 call. Coordinated response is another story. A peer reviewed study of the 2020 California fire season matched 7,214 wildfires against real aircraft flight data (ADS B): ``` Average aerial response time (CAL FIRE): 700 minutes (≈ 11.7 hours) Median aerial response time: 751 minutes (≈ 12.5 hours) Fires with response within 20 min: 55 of 7,214 (0.76%) Fires with NO aerial response in 24 hours: 5,737 of 7,214 (79.5%) ``` [Source: Magstadt et al., International Journal of Disaster Risk Reduction, Vol. 137, 2026.] That leaves a gap of about 699 minutes per incident between detection and aerial response. Nearly 80% of California wildfires got no aerial response at all in the first 24 hours. Tanith's target, and the brief's, is 90 seconds. That is the dent. ###### Why minutes matter. A 2024 analysis of California wildfire costs, published by the Moore Foundation, found: * Responding one hour sooner cuts the number of large fires by 16%. * Responding one hour later makes it 29% more likely that a fire burns more than 20 hectares. * Cutting average response by 15 minutes would be worth "billions" in economic benefit. [Sources: Dryad analysis citing Moore Foundation report; California Council on Science & Technology, 2025.] Tanith aims to turn a 700 minute response into 90 seconds. Even if real conditions stretched that to five minutes, it would still be about 140 times faster than today's average. Mati, seen through this data The Mati fire of July 23, 2018 killed 104 people and injured 172. Most died in their cars or homes in the first hours. In 2024, Greek courts found fire service and civil protection officials criminally negligent: drivers were sent toward the flames, evacuation orders came late, and the response was poorly coordinated. Survivor groups put the ruling plainly, as the 2025 anniversary coverage confirmed: "Most of the lives lost could have been saved." Not with better equipment. With a faster, better coordinated response. I'm not claiming Tanith would have saved 104 lives. It didn't exist, and that hypothetical can't be computed. But its architecture (AI detection, AI coordination, a human judge, a ground handoff in under 90 seconds) targets exactly the failure Greek courts spent six years documenting. The scale of the problem California alone had 8,213 wildfires with valid discovery data in 2020. The 2025 Los Angeles fires caused an estimated $164 to $275 billion in economic losses [Sources: PreventionWeb, UNDRR, arXiv]. California spends $1.13 billion a year on wildfire emergency response. Greece lost 104 people in twelve hours in one town in 2018. Australia lost 33 people and 3 billion animals in the 2019 to 2020 season. Portugal lost 66 people on one evacuation road in 2017. Climate models expect fires to grow more frequent and more intense through the 2030s and 2040s. If one regional agency, a CAL FIRE district or a Greek civil protection prefecture, adopted this architecture and closed even part of the gap, the impact would be counted in lives and dollars, not design awards. ###### What the dent looks like. The AI drone arrives first with intelligence. The manned helicopter follows, already briefed. The city scale endpoint. A protective layer nobody notices until it is needed. Post incident inspection. The legal audit layer becomes the tool for reviewing what happened. The system in motion. Dock to anomaly zone by the quickest path available. ##### Still being built. Tanith is a competition stage prototype with a full brand system, a live brand site, a deep research base and a clear architecture. The next steps are named: a pressure test with real dispatch operators, then a 30 day shadow pilot (the plan ). ###### The one thing I'd change. We did zero primary research with real disaster operators. The findings on operator behavior (Stack 3 ) came from published studies. They are good findings, but they aren't a week inside a CAL FIRE dispatch center. If Tanith were a real venture, that would be the next sprint. ###### Credits and contact. PROJECT Tanith, AI native command intelligence for autonomous emergency response COMPETITION FlytBase × Design Fusion Season 3, 2025 to 2026 DURATION 9 days, December 2025 to January 2026. TEAM 4 (Chaos Crew) Abhishek Paatni Lead. Brand, product strategy, IA, UX, Framer site. Vanshika Kandya UI screen execution. Saniya Jadhav Research support, secondary scans. Lakshmikanthraju HP Motion graphics, logo refinement. LIVE SITE tanith.framer.website CASE STUDY https://tanith-intelligence.vercel.app/ CONCEPT VIDEO youtu.be/9ZaovEzCepA EMAIL thepaatni@gmail.com PORTFOLIO paatni.vercel.app SCHEDULE topmate.io/thepaatni/1772841 (block a discovery call) Tools Figma, Framer, Adobe Illustrator, Notion, Perplexity, Midjourney, Canva, Claude, ##### Nine days. Four people. Whatever sleep was left over. The work as it actually looked: Google Meet windows at 11 PM, the silence after a hard call, waiting for someone's frozen video to come back. Day 1, 11:40 PM. Hands on faces, hands on chins. The first working call where the scope of what we'd agreed to started to feel real. Day 2, 12:49 AM. Lakshmi on a Krishna avatar because his camera had decided enough was enough. We kept going. Day 2, 7:44 PM. Presenting the gameplan, with the name list on screen: Tanith, Hedron, ORDIN. Three candidates. One survived. Vanshika, full focus. The UI screens in the Solution chapter went from sketch to shipped in her hands across four nights. Day 2, 5:07 PM. Strategy review. Three of four on the call, in front of bookshelves, a balcony, a blank wall. The work happens wherever it happens. Day 2, 8:27 PM. Heads down. Different rooms, different cities, one document open in everyone's tab. Day 4, 8:24 PM. Three of four, mid decision. The middle of a project is where teams either pull together or split. We pulled together. If you have done a competition like this, you know it doesn't look like the screens in the solution . It looks like these screenshots. Four people in four rooms, two hours into a call, arguing over whether the confidence threshold should be 85% or 80% while someone tries to remember if they ate lunch. Vanshika ran the screen execution under brutal time pressure. Saniya carried part of the research and helped work out the trust mechanics. Lakshmi held the motion and the final logo passes when there were no hours left. I led it, and I'm proud of how. I'm prouder of who was in the room. Up next Reframe & Role Back to portfolio ### Vinyasa (source: https://vinyasa-casestudy.vercel.app/llms.txt) #### Vinyasa case study: a gentler way to check in > Abhishek Paatni's case study of Vinyasa, a voice check in that replaced a long onboarding form: research, user journey, IA, edge cases and impact. * Author: Abhishek Paatni, Product Designer and Design Engineer (https://paatni.vercel.app/) * Case study: https://vinyasa-casestudy.vercel.app/ * The app: https://vinyasa-app.vercel.app * Published 2026-10-15, updated 2026-11-08 I joined Vinyasa in 2024 as its product designer. People weren't filling in our forms, and the team kept chasing them. I led the fix. This is the rebuild. ##### About Vinyasa Vinyasa helped people reach the right healthcare professionals, including therapists and psychiatrists. It was working on two problems: people being misdiagnosed, and people never being directed to the right mental healthcare. ##### Brief: Vinyasa's users weren't filling in its forms I was tasked with solving one problem: new users weren't finishing the onboarding questionnaire, however often the team followed up. ###### Project Overview * Client: Vinyasa, Divya Shah, Founder * My role: In house Product Designer and Builder * Year: 2024 * Duration: 21 days * Platform: Progressive Web Application (PWA) ###### What was going wrong Every new user had to fill in a long onboarding questionnaire before Vinyasa could help them. Almost every mental health platform works this way, BetterHelp included, so I understood why we did too. Then I read the therapy session transcripts and talked to people in therapy on Vinyasa. The form looked fine. Its format was the problem: people in real distress often lack the mental capacity to type out long answers. That was enough to convince the founder to try something else. ###### How I found this out I asked the founder for access to Vinyasa's therapists, users, session transcripts and team. Then I asked users to sit down with me, one on one. * I read the transcripts of the therapy sessions. To see what people were telling us in therapy. * I listened to the sessions, not just the words. One on one video calls with the same people, about what they were going through. * I heard what the words missed. Their voices carried feeling the transcripts never showed. The check in grew from that. ###### Three insights that shaped the design * Long forms ask too much of people who are already struggling. The people with the least to give were asked for the most. Every extra field was one more reason to stop. In Vinyasa: Questions you answer with a tap, and two short spoken tasks instead of an essay. * Voices carry what words miss. Transcripts showed me what people said. Their voices on our video calls showed me how they felt. In Vinyasa: I suggested we read emotion from people's voices, not only from their answers. * Trust has to be designed. People should know what a score means and where their voice goes. A calm interface isn't the same as a trustworthy one. In Vinyasa: Plain words for every score, a clear “not a diagnosis” label, no account and nothing stored. ##### Problem: forms lose the people who need help most Long forms caused people who were already mentally distressed to drop out of onboarding, so Vinyasa never got the signals it needed to direct them to the right help. More than 75% of new users dropped off during onboarding. That left users without help, therapists unable to support them, and the business struggling to move forward. How might we make a mental wellness check in easy enough to finish and useful the moment it ends? ##### Solution: a short check in that listens Vinyasa became a short mental wellness check in that uses audio, a person's own voice, to understand how they're doing. A few taps and two short spoken tasks replace the long questionnaire, so nobody has to type long answers. This rebuild goes one step further, into sonic therapy. After scoring the check in, it plays calm sound, from soft keys to rain, picked to ease stress and anxiety and give some rest and relaxation (R&R). ###### User journey map * 1. Arrive. Doing: Opens Vinyasa in a browser. Risk: Doesn't know what it will ask of them. Design answer: No account or download. “How Vinyasa works” is one tap away. * 2. Rate and consent. Doing: Rates their stress from 0 to 10. Risk: Worries about where their voice goes. Design answer: One plain sentence: Gemini hears it, Vinyasa keeps nothing. * 3. Check in. Doing: Answers 5 short questions. Risk: Typing long answers feels like too much. Design answer: One tap per answer, and Next waits until they're ready. * 4. Speak. Doing: Reads a passage aloud and describes a picture. Risk: Feels awkward, or the microphone is blocked. Design answer: Both optional, 30 seconds at most. Skip is always there. * 5. Report. Doing: Reads 7 areas in plain words. Risk: Takes a score as a diagnosis, or panics at a high one. Design answer: Labelled “not a diagnosis”. Very high scores point to a doctor, therapist or helpline. * 6. Listen and recheck. Doing: Plays a calm sound, then rates their stress again. Risk: Leaves with nothing to do next. Design answer: One sound picked for them, and the change in their own rating. ##### Research: what Vinyasa was able to uncover Vinyasa's view of mental wellness came from countless studies, research and the practical experience of its therapists. Together they showed that parts of mental wellness can be tracked and scored through sound. Human voices carry feelings and emotions, and a trained model can pick those signals out of audio. In the original MVP I used the Thymia API to do this and score users against standard mental health criteria. ###### The three studies that matter most * Music and stress: 104 randomised controlled trials, 9,617 participants, Psychological stress outcome: d = 0.545. Music eased stress, more in how people felt than in their bodies. d = 0.545 is an effect size, not a 54.5% drop. Source: de Witte et al., 2020, Health Psychology Review, https://pubmed.ncbi.nlm.nih.gov/31167611/ * Music therapy and stress: 47 studies, 2,747 participants, Outcomes related to stress: d = 0.723. Music therapy, a structured approach, had a medium to large effect. An app playing music is not the same thing. Source: de Witte et al., 2022, Health Psychology Review, https://pubmed.ncbi.nlm.nih.gov/33176590/ * Why results vary: 14 experimental studies, 706 participants, Overall effect in the short term: not statistically significant. Lab studies found no clear overall effect of music on stress recovery, so Vinyasa never promises one. Source: Adiasto et al., 2022, PLOS ONE, https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0270031 ###### More findings, with their sources * 98% of stress related adjustment disorders go undetected. GPs detected only 2 of 110 cases of adjustment disorder, a diagnosis linked to stress. 3,815 primary care patients, Catalonia. Source: Fernández et al., 2012, British Journal of Psychiatry, https://pubmed.ncbi.nlm.nih.gov/22576725/ * 71% of distress in cancer patients goes unrecognised. Doctors recognised only 29% of patients whose GHQ 12 scores suggested psychological distress, so about 7 in 10 went unrecognised. 2,297 patients, 34 UK cancer centres. Source: Fallowfield et al., 2001, British Journal of Cancer, https://pubmed.ncbi.nlm.nih.gov/11308246/ * 86% of burned out teachers also meet criteria for depression. 86% of teachers classed as burned out met criteria for a provisional diagnosis of depression on a standard screening form. 1,386 US teachers. Source: Schonfeld and Bianchi, 2016, Journal of Clinical Psychology, https://pubmed.ncbi.nlm.nih.gov/26451877/ * Sleep: Music improved sleep as adults with insomnia rated it, but not clearly when sleep was measured. Source: Jespersen et al., 2022, Cochrane, https://www.cochrane.org/CD010459/music-listening-to-improve-sleep-quality * Depression: With a qualified therapist, music therapy added to usual care eased depression in the short term. Source: Aalbers et al., 2017, Cochrane, https://www.cochrane.org/evidence/CD004517_music-therapy-depression * Mental fatigue: In small experiments, upbeat music helped with mental fatigue more than relaxing music did. Source: Ding et al., 2025, PLOS ONE, https://pmc.ncbi.nlm.nih.gov/articles/PMC11698372/ * Natural sounds: Natural sounds lowered some body measures of stress, but people didn't report feeling less stressed. Source: Fan and Baharum, 2024, Stress, https://pubmed.ncbi.nlm.nih.gov/39285764/ * Voice: In studies, voice patterns help tell apart speakers with and without depression. It's a signal, not a verdict. Source: Briganti and Lechien, 2025, Journal of Voice, https://orbi.umons.ac.be/handle/20.500.12907/53444; Martínez Nicolás et al., 2025, Frontiers in Psychology, https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2025.1715538/full * Drop off: Across 93 mental health apps, about 4 in 100 users were still active after 15 days. Source: Baumel et al., 2019, JMIR, https://www.jmir.org/2019/9/E14567/ * Short forms: A short PHQ 9 form of three questions had 98% sensitivity and 76% specificity, not yet checked against interviews. Source: Regenstrief Institute, https://www.regenstrief.org/article/shortening-the-patient-health-questionnaire-9 * Seven areas: Vinyasa's seven areas are its own. They cover similar ground to common questionnaires but aren't one. Source: Novopsych, https://novopsych.com.au/assessments/depression/depression-anxiety-stress-scales-short-form-dass-21/ ##### Prototype: the rebuild, two years after the original ###### From the 2024 MVP to this rebuild The original MVP, built in 2024, ran on the Thymia API and a HIPAA compliant language model. It was later acquired by a US mental health startup. Two years on, I rebuilt it from scratch and made it free to try. Every screen below is from this rebuild. It runs on free tools, so its voice readings are less accurate than the original's. ###### Information architecture * Welcome: First name, if they want; How Vinyasa works; Menu: privacy, terms, case study * Check in: Stress rating and voice consent; 5 questions, one tap each; Read aloud (optional); Describe a picture (optional) * Results: Report: 7 areas; Scorecard to save; Your next step: one sound * Listen: 22 calm sounds, made in the browser; Stress rating again; Check in again, or finish * Need help now? Helplines, one tap away on every screen. * Crisis language If the voice step hears it, helplines come before the report. ###### Six Moments in the User Journey * Consent and baseline: A 0 to 10 rating of how stressed you feel, and a note on your voice. * Quick check in: Tap an answer, then press Next. Skip is always there. * Read aloud: An optional reading of up to 30 seconds. Skip is one tap away. * Describe a picture: Describe a calm picture in your own words. Skip anytime. * The report: Seven areas in plain words, all from your own answers. * Your next step: One recommended sound, with other sounds a tap away. ###### Edge cases The rebuild runs in a browser, on free tools and a tight timeline. Those limits decided what I handled and what I traded off. Accounted for: * Microphone blocked or missing: A plain message, and Skip still leads to a result. * Silent or very short clip: Caught before anything is sent. Try again, or skip. * Slow or failed voice analysis: Two AI models race. If both are slow, the report uses the answers alone. * An unexpected AI reply: Checked against a strict format. The AI's own words are never shown. * Crisis language: Helplines come first, and the report stays one tap away. * Skipped questions: Scored from what was answered, with “Not enough to say” for the rest. Traded off: * Crisis phrases (Time): English and basic Hinglish only, not yet reviewed by a clinician. * Voice accuracy (Budget): Gemini's free tier is less accurate than the Thymia API the 2024 MVP used. * Free quota (Budget): When it runs out, the report falls back to the answers alone. * Locked iPhone (Platform): iPhone may pause browser sound when the screen locks. The app says so. * Screening needs a voice (Platform): A skipped voice step can't be screened, so “Need help now?” stays one tap away. ##### Outcome: what changed, and what a session leaves ###### The impact of the 2024 MVP * Before: More than 75% of new users dropped off during the long onboarding questionnaire. * The redesign: About 3 minutes for the check in that replaced it: a few taps and two short spoken tasks. * The build: 21 days to design and build the MVP. * The result: Acquired by a US mental health startup. ###### Before, during and after Screens from one session made with sample data, not a real user. * Before: A 0 to 10 stress rating before listening sets the baseline. * During: A calm scene and simple controls keep the sound in front. * After: One more rating shows the change. Your own rating, not proof. ###### What I took from it A check in only works if it asks very little. A product only works if people trust it. Both came from listening before I designed anything. This rebuild doesn't claim to fix mental health issues or act as a treatment. It's a free rebuild of the original MVP that lets people check in on their mental wellness, so you can try the solution yourself. Vinyasa is a wellness check in, not a medical service. If you're struggling, talk to someone: India's Tele MANAS, 14416. ### Start Anyway (source: https://startanyway.vercel.app/llms.txt) #### Start Anyway > Start Anyway is a free Android app built around a companion that gets you to start. You pick one of four companions and name it, tell it your big goals by voice or by text, and it turns them into three plans and then into small tasks for today, sized to your time and your energy. You press and hold a key to start a task, your companion takes over the screen while you focus, and it calls you out when you scroll too long. It installs from a file on this website because it is not on Google Play yet. ##### What it is * An app for people with a big goal that feels too heavy to begin, and for anyone who loses evenings to scrolling. * Four companions, each with its own personality, voice and way of moving: Kade (strict), Juno (soft but realistic), Orin (logical) and Nori (soft). You name yours, and you can change it or rename it in Settings. * You say or type up to three goals. Your companion reads them back as cards you can edit, remove or say again. * The AI writes three plans: All in, Real life and Hard season. Each has minutes a day, days a week, a time estimate per goal, three milestones and the week's focus. * Each day you give your time, your body energy (1 to 5) and your mind energy (1 to 5), and get three to five tasks that fit. Low body energy means nothing hard on the body; low mind energy means simple work. * Hold to start: you press and hold a slanted key for about a second. A cut in band slashes across, and your companion takes the whole screen for the focus session, with a timer. Focus mode can pin the screen with Android screen pinning; an X at the top left always lets you break the session, and the minutes still count. * The brain rot watch, off until you turn it on: when an app you picked is open, your companion drops in from the top of the screen like an island. It speaks up at 15, 20 and 25 minutes, and at 35 minutes it says you don't seem serious about your goals and leaves. * Free. No account. No ads. No analytics. ##### Platform and status * Android only. Not announced for other platforms. * A direct download of an APK file from the website. Not on Google Play yet. ##### Data * Goals, plans, history and the companion are kept on the phone, in one file in the app's private storage. Android backup is off for the app. * When you set your goals and each time you ask for today's tasks, what you told the app about your goals, your plan and your day is sent to Google's Gemini AI, on its free tier, through the Start Anyway service on Vercel. If Gemini is busy, Groq answers instead. Google may use free tier text to improve its products. * The brain rot watch only knows which app is open and for how long. Which apps you use stays on the phone and is never sent to the AI or anywhere else. It never sees the screen. * Delete all my data is in Settings. ##### Pages * Home: https://startanyway.vercel.app/ * Privacy Policy: https://startanyway.vercel.app/privacy.html * Terms and Conditions: https://startanyway.vercel.app/terms.html * Data and Security: https://startanyway.vercel.app/data-security.html ##### Founder * Built by Abhishek Paatni, a designer who wanted a way to turn long term goals into the next small step. ### unfold (source: https://unfoldnow.vercel.app/llms.txt) #### unfold > unfold is a free two player question game for one Android phone. You pick a vibe, flick a die, take turns answering out loud, swipe "ate" or "veto" on each card, and the winner gets a story sized receipt poster. No account, it works offline, and it never asks for internet permission. Made by Abhishek Paatni, a Product Designer and Engineer. ##### Facts * Platform: Android 10 or newer, installed from an apk on the home page. Play Store and App Store versions: coming soon. * Price: free. There is nothing to buy inside the app. * Version: 2.4, updated 28 October 2026. About 10 MB. * Players: 2, sharing one phone. * Content: casual profanity and flirty questions. The "unhinged" question pack is 18+ and the app asks both players to confirm. Intended for ages 16 and up. * Permissions: vibration only (for haptics). No internet, camera, storage, location or contacts permission. * Data: nothing is collected. Names, answers and scores live in memory and vanish when the app closes. Only the sound, haptics and calm motion settings are saved on the phone. ##### Pages * [Home](https://unfoldnow.vercel.app/): what unfold is, how it works, answers to common questions, install steps and the download * [Privacy Policy](https://unfoldnow.vercel.app/privacy): what the app and this website do (and do not do) with information * [Terms and Conditions](https://unfoldnow.vercel.app/terms): who can use it, the free personal licence, and the usual limits * [Data](https://unfoldnow.vercel.app/data): where every piece of data lives, for how long, and how to delete it * [Download the apk](https://unfoldnow.vercel.app/unfold.apk): the Android installer ##### Contact * thepaatni@gmail.com ## Contact and links * Email: thepaatni@gmail.com * X: https://x.com/thepaatni. My design profile: skills and craft, and a flamboyant designer. * Instagram: https://www.instagram.com/paatnithisside. Personal. No business in my DMs. Connect here as a fellow designer. * LinkedIn: https://www.linkedin.com/in/thepaatni. Professional only: what I know and whatever I’ve learned. * GitHub: https://github.com/paatnithisside * Book a call: https://topmate.io/thepaatni/1772841