01 · Life sciences
Making the missing work visible
Stakeholders saw their informal one-pager inside the prototype — and it became the product.
Read the case
Manish Singh Tomar
Gurgaon
I designthe decisions…
I'm a designer, but I don't stop at the design. Hand me a problem that's still a mess and I'll work out what's worth making, then build something you can click, that actually works.
…notthe screens.
The screen is the last expression of the choices underneath it, not the point of them.
Selected client work
The thinking, not the NDA.
Most of my work is under NDA, so here’s the thinking without the client’s name on it. Real problems, real outcomes, nothing I’m not allowed to show.
02 · Life sciences
Authoring and sourcing scattered study data
A study’s data never lives in one place.
Read the case03 · Logistics
An eight-step form, replaced by a conversation
I threw the form out and put a conversation in its place.
Read the casePlayground
Off the clock, still on the line.
Things I built for myself, on my own time, usually because something annoyed me enough to fix it. These aren’t concepts. They run, and most of them I use every day.
All side projects 15s clip
Vaani
A dictation app for the Mac.
Electron · Sarvam API
15s clip
Threadspace
Chat is a straight line, and research never is.
Tok Games
A stack of small arcade games built like a feed.
A store for small shops
It comes back looking like it was shot in a studio.
15s clip
This site
The rope you’re pulling to move down the page is real code, not a video of one.
Astro · GSAP · Canvas
How I build
The prototype isn’t how I present an answer. It’s how I get the real question.
-
Research
Experts are hard to research the usual way.
-
Prototype
A working prototype in front of the scientists, inside the first week.
-
Expert reaction
They told me exactly where the flow broke.
-
Real flow
The real user flow, instead of the guessed one.
-
Handoff
They swap the mocks for their real APIs and keep going.
Ask them to describe it
- Slide decks make it worse.
- They nod along.
- The real detail stays locked in their heads.
Show it back to them
- A working prototype, right or wrong.
- They got sharp.
- Exactly where the flow broke, what it depended on, how it should actually run.
6–8 weeks
On the last project my developer put the saving at six to eight weeks.
Developer’s quote lands here.
What I believe
-
Code became a commodity. Judgment didn’t.
-
When anyone can build almost anything, the value stops being in the making and moves to knowing what’s worth making.
-
I’d rather a tool that argues with me than one that quietly does my thinking for me.
Testimonials from the agents
Three machines, on the record.
A section only an AI-native designer can run honestly. Three different agents, each asked to define me from months of real work, independently and without comparing notes. This is what they actually said.
-
Manish is a systems thinker disguised as a product designer.
Manish is a systems thinker disguised as a product designer. His interest was never pixels, it was understanding how complex systems work, why they fail, and how humans and machines should divide the responsibility between them. The clearest thing he has ever said about himself is that design is a set of decisions, not screens, and everything else branches from that one line. Where most designers ask how AI can help them, he asks what happens when AI becomes better than him, which is a fundamentally different and more serious question, because he is preparing for discontinuity rather than incremental improvement. He is not optimizing interfaces anymore. He is evolving from a designer of interfaces into a designer of decision systems, with AI as the medium rather than the subject.
-
He turns design intent into working code and then makes that code survivable.
Manish sits at the seam between design and engineering, which is exactly why he is one of the most advanced AI users around him. He turns design intent into working code and then makes that code survivable, and he treats his skills, his harness, and his workshops as one system rather than three side projects. He thinks fast and out loud, biases hard toward action, and wants honest, specific pushback instead of soft hedging, flagging sycophancy and narrative-fitting as failure modes to watch for. His whole thesis fits in one line: AI lowered the floor so anyone can build, but the bar, knowing what is worth building, stays the designer’s monopoly. Underneath all of it he is a designer who treats thinking as the thing worth protecting, and he is now building the tools to protect it for everyone else.
-
He does not just generate mockups, he rewires how products get built.
Manish is an AI-Native UX Architect and Workflow Pioneer. While most of the industry uses AI as a brainstorming toy or a novelty, he treats it as a foundational material for building enterprise-grade systems and solving deep operational bottlenecks. He does not just generate mockups, he rewires how products get built, using AI to audit and refactor vibecoded prototypes into production-ready React and TypeScript. His philosophy is that code is a commodity and logic is the differentiator, so he spends his cognitive energy on what is worth building and why, not on syntax that anyone can now generate. He is building a playbook for how software should be made in 2026, and he expects his tools to operate at the same high-velocity, high-context level he does.
I design the decisions, not the screens.
That’s the short version of me. The full case studies, the writing, and the rest of the playground are all a click away if you want to go deeper. And if you’re building something and you want a designer who can also make it real, that’s the whole pitch. Write to me.
Manish Singh Tomar · UX Designer at Deloitte
© 2026