Product Designer — AI products, design systems, 0‑to‑1
A Product Designer with a CS/ML background. I take AI products from user research through to a launched, polished interface — interaction, design systems, and the build itself.
Open to Product / UX Designer roles · Noida, relocating, or remote · 30-day notice
Three projects · Ad-tech, veterinary healthcare, fintech
Ad-tech · B2B SaaS · AI-native
Agentic campaign assistant for a retail-media DSP
Launching across 12 networks took 1.5 hours of manual setup. The call: let AI draft the whole campaign, but gate every section behind human review — so speed never cost trust.
3× faster launch · 4.6/5 usability · −8% draft abandonment
Read the case study
Veterinary healthcare · Enterprise CMS
0-to-1 CMS and booking across a 130-clinic network
A 130-clinic network each wanting a bespoke site. The call: standardize to 6 templates instead — the trade-off that made the whole network maintainable without gutting local identity.
+41% organic traffic · +16% booking conversion
Read the case study
Consumer fintech · Concluded before build
Conversational finance app for iOS
Finance apps lead with net worth — a number that answers nothing about today. The call: one Safe to Spend figure that always shows its arithmetic, and two named voices so reassurance and opportunity never sound the same.
Full process, whiteboard to hi-fi · Stopped at v1 scope, before engineering
Read the case studyFigures are internal metrics measured post-launch.
Selected design work
Interfaces, brand systems, and marketing design. Scroll through — click any shot to enlarge.
What people I've worked with say
“He is the person who, if he commits, delivers with utmost excellence.”
Suman worked on our dream project as a Lead. His gift to perfect the designs was seen from the very beginning — from understanding the founder's vision to executing the minutest details.
His sincerity and dedication to execute plans despite numerous hurdles is what makes him a true winner. I believe his grit will carve his mark in this industry.
Samira Pujari
Founder, RaikarsGoodVibes — art & décor brand strategy
FAQ
The questions recruiters and hiring managers usually ask first — answered up front, so we can skip straight to the work on our first call.
End-to-end product designer. I run discovery and user research, design the interaction and visual layers, build and maintain the design system, and ship the front-end in React/Tailwind. AI-native design is my focus — agentic and conversational UX that models can actually serve.
On Carter I designed an agentic AI campaign assistant on the Claude API that drafts a campaign from a brief, with human review of every section and per-network spec checks — the clarification, confirmation, and recovery patterns that decide whether an agent feels trustworthy or just fast. Campaign launch went from 1.5 hours to 30 minutes. On Poppins I designed a conversational finance assistant built on two depths of a single voice, each with separate jobs and an explicit list of what it never says, because the tone that works for "you hit your savings goal" is unbearable attached to "you are overdrawn." The part I care most about is the unglamorous part: what the interface does when the model is wrong, uncertain, or slow.
Carter (B2B ad-tech) — design across 12 DSP integrations as one of two designers, an agentic AI campaign assistant, and a multi-DSP launcher unifying 12 workflows into one form; 4.6/5 usability. Atlas (veterinary care) — a 130-clinic network unified under one relational CMS with 6 WCAG 2.1 AA templates and a 0-to-1 no-code site builder; booking conversion up 16%, organic traffic up 41%. Poppins (consumer iOS) — a conversational finance app taken from concept to hi-fi prototypes, then concluded before build when the business case didn't hold. Research, interaction design, and UI on all three.
Because we couldn't defend the wedge. We'd bet on proactive delivery — money surfacing in the messaging thread people already read. Under scrutiny that didn't hold: a delivery channel is a feature, not a moat, and cheap for a funded incumbent to copy once it's proven. Everything genuinely differentiated lived in voice, sequencing, and interaction quality — real design value, but a fragile business defence. We concluded it before build, so no engineering was spent. What I'd change is the order: we validated the experience exhaustively and the wedge late. Now I go at the assumption that kills the project first.
Product Designer or Senior Product Designer, on scope rather than title. In two and a half years I've taken a 0-to-1 veterinary-care CMS across a 130-clinic network from nothing to launch, led design across 12 DSP integrations on a live retail-media platform as one of two designers, and built the company-wide design system now used by five products. I've also taken a stalled-draft problem to leadership with interviews, session recordings, and SQL rather than taste, and got it prioritised ahead of the roadmap. Read one case study and judge the decisions — that's a fairer test than the date on my CV.
Product Designer or Senior Product Designer somewhere design owns the problem from research through shipped UI, and AI is part of the product rather than a bolt-on. Company size matters much less to me than that — I've worked small enough to talk to every user directly, and inside an 80+ person org. Bonus if I sit close to engineering.
All three are on the table. Based in Noida, India, and open to relocation for the right role. Happy to overlap with US or EU hours — the GoDiverse engagement ran fully remote with a UK team.
I'm an Indian citizen based in Noida, so roles in India need no sponsorship. For the US, UK, or EU I would need it — better to raise that in the first conversation than to find it at offer stage.
30 days. I'm on a standard notice period and I'll honour it — a clean handover matters. If the role is right, I can start knowledge-transfer conversations immediately.
Talks & workshops
Three design workshops for students — colour theory, Figma, and taking an idea from a blank canvas to something you can click. Each for 150–200 attendees, mostly full computer labs working along live. Explaining a decision to a room who has never made one is the fastest way to find out whether you actually understand it.
