Product strategy · Design · AI agents · Engineering
We help startups build better products, engage users, and accelerate growth through AI-first strategy, design, and engineering.
Where you are
Wherever you're starting, bring in the capabilities you need to move forward.
How we help
Wherever you are starting from, the way forward runs through five connected stages. Bring us in for the stage you need now, or for the whole journey.
Find direction
We validate the opportunity, learn what your users need and help you decide where to invest first. You come away with a clear product direction, a roadmap your team can start on and a shared understanding of what to build.
A good first step when you have an idea and want a plan to build it.
Design the system
Product design and user research shape experiences that feel consistent and easy to use. Design tokens, reusable components, design rules and production code come together in one shared foundation that your team and your AI tools can build from, in Figma and directly in code.
Useful when your designs and your code need one source to work from.
Build and ship
We design, develop and deploy custom platforms and software. We can work alongside your engineers or take full ownership of delivery, from the first concept to production, on foundations ready for growth. The software and the code belong to you from day one.
For products ready to be built, finished and launched.
Put AI to work
We build practical AI agents and intelligent workflows, connected to the systems you already use. Each one runs on the model best suited to its task, keeps a person in charge of the decisions that matter and reports results you can measure.
For teams ready to bring AI into their product and their daily operations.
Learn and evolve
Research, analytics and experiments show where to improve next, from conversion and activation to retention and operational performance. We help your team spot the opportunities, prioritise the changes with the most impact and keep the product developing.
For live products ready to grow their numbers.
Take one stage or the whole journey. Each stage builds on the one before.
Most of it, yes.
Surprisingly far…
…then a wall.
What it can't do.
That's the part that eats your week.
Who you're talking to
Himanshu Gupta. 15 years building products, not advising on them.
Six years at PharmEasy, from site launch to 30M+ users and 200K+ daily transactions, running 30+ people across 24 products. Then Head of Design and Research at Infra.Market.
In 2024 I left to build Therapon, an AI clinical assistant. Retrieval across DSM-5 and ICD-11, automated assessment selection, and an eval harness, because a wrong answer in a clinical setting isn't a bug report. Clinically validated with Montfort University. I've been building agents and workflow automations ever since. So when we talk about what your agent should decide, where the human stays in, and how it fails safely, that's from shipping it, not reading about it.
I'm back to consulting because what founders kept asking for wasn't strategy decks. It was the missing capability: design, product and engineering, on demand.
No account manager. No junior swapped in after the sale.
The team
Disciplines:
I pick the specialist for the problem you actually have, and I lead the work throughout.
On your terms
Choose the way of working that suits your team today, and change it as you grow.
We take your product from first idea to launch and stay accountable for the outcome.
We join your designers and engineers and work in your tools, your rituals, your codebase.
A senior specialist sits inside your team, in your standups and your repo, for as long as you need them.
After launch, we keep improving, maintaining and scaling the product with you.
Weekly demos, not status decks. You see the product, not a report about it.
A ship date, written down before we start.
One month's notice. No lock-in, no minimum term.
You own everything on final payment. Code, files, prompts, documentation.
If I think you should hire instead of engage me, I'll say so on the call.
Credentials
Our team has built and scaled startups at Microsoft, Flipkart and PharmEasy — and now builds for founders doing the same.
Where we've built
Clients
Sectors
AI agents
Four kinds of agentic work. Yours will look different.
The copying your team still does by hand.
Fastest payback, usually within a month.
One repetitive job, the same way every time.
Consistency as much as speed. Auditable.
Support, onboarding, qualification. Grounded in your content.
Highest ceiling, slowest payback. Do the internal work first.
Evals, cost control, drift.
The difference between a pilot and a product.
Therapon
AI clinical assistant
3 minCASE PREP TIME
85%TIME REDUCTION
Emami Frank Ross
100-year-old pharmacy, online
PharmEasy
Healthtech · 0 to $5B
30M+USERS
200K+DAILY TXNS
AVP Design & Research, 2017–2022.
Schedule a Strategy Call →Reviews
"Himanshu's leadership in building and scaling our MVP for millions of patients and caregivers has been pivotal. His deep understanding of user needs has shaped everything…"
"I've worked on several products with Himanshu, all of which found commercial success. He remains a pioneer in the product and design landscape in India."
Book a call
FAQs
Engagements are priced monthly and scoped to the work. Most run between $5,000 and $15,000 a month. You get a defined scope and a named end date before anything starts, and you can stop at the end of any month.
Usually, and often not where people expect. The fastest return is internal: the work your team does by hand moving things between tools. Customer-facing AI gets the attention and takes longer to pay back. If you already run AI in production, cost usually comes down through model routing and caching rather than a rebuild. Either way, we scope against a number you already track, so you can tell whether it worked.
Products change while you're building them. Monthly means the team scales to what the work needs that month. No lock-in, no minimum term.
Depends what “it” is. A working prototype or a single agent doing one job can be live in 48 hours. A full product is four to eight weeks from a defined scope. Either way the date goes in writing before work starts.
I lead every engagement, personally, throughout. Specialists in AI, product, engineering, research and design come in for what your project needs. No account manager, and nobody substituted after the sale.
Both. Scoping, flows, tool connections, evals and the build. Most of the time goes into what the agent is allowed to decide and where a human stays in the loop, because that's where these systems fail.
Yes. I'll review them first and tell you honestly if anything should change before build. Fixing something in Figma costs a day. Fixing it in code costs a fortnight.
You do, all of it on final payment. Design files, source code, prompts, documentation, everything. It's written into the agreement, not assumed.
Ask me on the call. Thirty minutes, and you'll leave with a scope and a timeline whether or not we work together.
Capabilities
No templates, no stock, no bought assets. Open any card to see it running.
Your operation as a place you can walk around, with your real data in it.
Five seconds: the problem, the hook, the fix, the payoff.
Campaign assets that look made, not generated.
The screens, the flows, and the decisions behind them.
Interface motion that explains rather than decorates.
Web and mobile, shipped and running.
Agents, automations, and the systems that keep them honest.
Thirty minutes. You leave with scope, timeline and a view on cost — whether or not we work together.
