Your model is right. Your users still won't bet on it.
That half-second of doubt is where AI products go to die — and it's the only thing I design. Confidence that ends in an action. Reasoning a person can interrogate. An override the system learns from. Fifteen years on this one problem; every number below arrives holding its baseline.
I read the model at the eval layer. I shape what gets measured. I ship the front-end. · Founding · Staff IC · Director-level · Available · 4 weeks' notice
Arpit Maheshwari
Design leader · AI & data-intensive products
By Friday of week one I've read your evals and sat in your customer calls; by launch, the interface I drew is code I shipped. Indore, India · fully remote · 4–5 working hours shared with US East every day.
Selected Work
Fifteen years of this work mostly lives behind NDAs. The six below are the shape of all of it — two told in full, four as recorded decision walkthroughs you can watch right now.
PTC University — Learning Connector
Five learning platforms, one survivor, 11 languages. Drawing the screens was the easy part; convincing a company to retire four products and rewire its revenue model was the design work that mattered. Subscription went 0% → 64% of new bookings inside a year; print costs fell $1M/yr.
Telefónica MyO2 & Priority Moments
Two O2 UK products at national scale, before "AI designer" was a job title — MyO2 self-service (4M+ users) and Priority Moments loyalty (2.6M sign-ups). Every screen drawn by me, then coded by me, for mobile web. The origin story of everything else on this page.
Four more · NDA · Walk-through on request
Programmatic Advertising Platform
Traders watched an algorithm beat them on the scoreboard and still played their own hunches. The fix wasn't a better model — it was a score tied to one action, reasons named out loud, and an override that taught next week's predictions.
AI-Assisted Private Equity Investing
Analysts get paid to doubt confident numbers, so I held the launch until the score could argue its own case. Once it could, they stopped auditing it and started leaning on it.
OrgOS · Transparent Org Tooling
Eight modules doing the coordination work a management layer normally does. The brief had no precedent I could find: the org chart and the user base were the same two hundred humans.
Technical Due Diligence Platform
Partners stake millions on technical claims they'll never personally verify. Extracting the signals was the model's job; getting partners to lean their reputation on the extraction was mine.
Patterns
AI Design Patterns Library
Everything in this library ran in production, failed somewhere specific, and came back stronger. The tradeoffs are written down because I paid for them once already — you shouldn't have to.
The Act / Review / Ignore rule
One score, one action. An unexplained 87% is a shrug with decimals — it tells the user how the model feels and nothing about what to do. Every confidence surface I ship resolves to exactly one of three verbs: act, review, or ignore.
Read the rule →Confidence Score Patterns
Five ways to put a number on certainty — numeric, color, language, composite — and when each one earns or burns trust.
AI Failure States
What the screen says when the model can't deliver, and how saying it honestly keeps users from leaving.
ML Explainability
Showing non-technical people why the machine decided, at a depth they can use without a stats degree.
Human-in-Loop Patterns
Keeping the person in command when the stakes outgrow the model's confidence.
How I Lead
Hire me and week one looks like this: I'm reading eval results before opening a design file, sitting silent on customer calls, and writing the diagnosis nobody assigned. By week two we're arguing productively.
What you're hiring me to own
- The trust layer
- The exact pixels where a person decides the model deserves their click, or doesn't.
- The design language
- A component system the next designer can run without me in the room, because it's documented, not memorized.
- The ML/UX contract
- A written promise between the model team and the user: here's what this system can do, here's where it taps out.
How I partner with each function
- With engineering
- Eval design before interface design, always. The CSS I own ships under my name in the PR. When we disagree on feasibility, the cheapest experiment goes first and settles it.
- With product / founder
- I'll contest the roadmap when the numbers contradict it, and I sign up for outcomes rather than deliverables. The design doc is mine to write; the spec is yours.
- With customers
- Five calls in my first week, one a week forever after, and I read the raw support tickets myself. No AI feature ships until I've personally watched someone fail to use it.
The errata
I once fought hard for a recommendation card with three ranked options — give people choice, I argued. Engineering wanted one option, the top pick, nothing else. They won the meeting. Next quarter's A/B test made it permanent: the single-option card converted 2.3× better, because three choices froze people at the exact moment we needed them to move. Every recommendation surface I've designed since starts from that loss.
Three things I'll refuse
- Being the only designer in a company past 40 people — beyond that point design is an organizational problem, and one heroic hire is the wrong answer to it.
- Shipping an AI feature with no designed failure state — this one isn't negotiable, and if that's a dealbreaker we've saved each other an interview loop.
- Running design ops while also shipping product — it's two jobs, and doing both is how senior designers quietly do neither.
What the people I've worked with say
Over the past four years at Talon, Arpit has been instrumental in shaping four distinct products from the ground up. His user-focused designs are remarkably intuitive yet adept at handling complex workflows… If you need a designer who excels at combining strategic vision with practical execution, Arpit is the person to call.
Arpit has worked with me for years and I value his honesty and hard work. He's been an integral part of my staff… involved in all facets of the team, from design to development to hiring and onboarding of new members.
Arpit teams up with designers very well, not only does he flawlessly execute the UI implementations but he pushes back on design decisions using his UX expertise… I'd recommend Arpit to any team looking to improve their final product.
About
Fifteen years across EdTech, telecom, AdTech, fintech, and org tooling — the last six of them fully remote, three companies, millions of users, zero shared offices. Home is Indore, India. I publish The Trust Layer, a newsletter on making AI products people actually act on.
On my first accessibility project I spent a week with my monitor switched off, navigating by screen-reader — designing for someone who isn’t you starts by becoming them.
Writing
The Agentic MVP: Why Your Next Launch Will Be Lovable, Not Just Viable
Every founder knows the pit in their stomach on Launch Day. How the rise of agentic systems is rewriting what "minimum viable" means — and why lovability is now the bar.
The AI Fight Club: Weaponizing Claude and Gemini for Bulletproof Products
Pitting AI systems against each other to strengthen product robustness. A practical method for stress-testing your AI features before users do it for you.
The New Renaissance: How AI is Transforming Us from Software Operators to Digital Artisans
Ushering in a new era of digital entrepreneurship. How AI is changing what it means to build, and what designers must understand about the tools reshaping the industry.
Contact
One seat. Full-time. Yours to offer.
I'm looking for exactly one role: founding designer at an AI product company of 5–40 people — or a staff / director seat where the trust layer is the actual job description. Fully remote from GMT+5:30 with 4–5 hours of daily US East overlap. Available — four weeks' notice.
Prefer to talk first? Book 30 minutes directly ↗ — I'll arrive having already used your product.
The skim version: one page — positioning, receipts, references, availability →
Not hiring but building something in AI? The patterns library and the founder checklist are free — take them.
What it's like working with me
Indore (GMT+5:30) runs ahead of your morning: overnight progress, then 4–5 shared hours every afternoon US East. Decisions get written down the day they're made — six years remote taught me documents outlive meetings. And I don't throw designs over a wall; I open the PR. Three companies' worth of shipped front-ends, millions of users, say so.