About
I am a builder who thinks like a product manager.
There is a very thin line between building the thing right and building the right thing, and you have to know which one you are on. Prototyping is how I find out, and it costs less than a retro.
How I work
Build the thing badly and it shows up fast, so it gets fixed. Build the wrong thing well and it does not show up for a quarter, and by then it has eaten the sprint that made it. Prototyping is the cheapest way I know to catch that. Put something clickable in front of the people who would actually use it, and the argument stops being about taste.
- Write the problem down first. If it can’t survive a paragraph, it won’t survive a sprint.
- Decide what a change is meant to move, and how you’ll know, before you decide to build it. The pre/post comparison should be designed with the feature, not after it.
- Stay close to the operational reality. Spacez Captains existed because the people using it were on WhatsApp at 11pm, and you only learn that by asking.
Now
I own retrieval and booking on the storefront at Spacez, plus the tools behind 50+ properties. Most of it comes back to one question. Someone arrives with a vague idea of a trip: a workstay, a group weekend, somewhere with a pool. How much of the distance between that and a confirmed booking can we actually remove?
Lately that has meant retrieval. I owned the requirements and the retrieval logic for a RAG-based villa assistant, combining BM25 with semantic search so property questions, offers and ticket escalation resolve in one place. A pre/post comparison put search-to-booking conversion up 15–20% relative. Pricing and coupon experiments moved website bookings from 5% to 35% of total, and past 45% of revenue. That is 7× the direct share, and really a story about depending less on the OTAs.
The other half is the unglamorous half. Check-in, check-out and property updates now run through n8n, which took about half the manual handling out of guest communication.
What I build
Agents, text and voice. The voice ones are what I am on at Spacez now, and building what I am specifying keeps a spec from being a guess. The rest started as a question I wanted answered rather than a product I planned.
Review Radar came out of watching teams spend 50+ hours a month hand-sorting feedback. It groups app-store and Reddit reviews and pushes prioritised summaries into Slack, cutting a digest from 50 minutes to about 3.5. Eleven PMs and designers rated the output 4.4/5, and 73% said they’d trust it for sprint planning.
Market Intelligence System is a daily brief for founders and product teams, ranked and summarised through n8n and Gemini, and delivered over Telegram in under two minutes.
How I got here
I came to product sideways. B.Tech in Civil Engineering at IIT Roorkee, 2022–2026. Four years of being told that a structure is only as good as the assumptions under it, and that you should know which load is actually governing before you size anything. I have stopped apologising for it. That transfers.
I joined Spacez as a product intern in June 2025 and spent it on operations rather than the storefront. First Spacez Captains, the app for the people running properties on the ground. It went live across 50+ properties and took average check-in from 30 minutes to 8. Then Spacez One: Console, where 7,800+ historical bookings and the OTA channels finally landed in one system. I stopped being an intern in January 2026. The dates and the title are on the résumé.
Before any of it, I co-convened Civil Conclave on campus: a 30-person team, registrations up 50% to over 300, INR 4 lakh from 10 sponsors. First real lesson in how much of a plan is just getting people to agree on one.