I started writing code at Samsung Electronics, building DRM modules, managing security on Tizen, and filing patents in India and the US. Then IIM Kozhikode taught me why products succeed or fail.
Now I sit at the rarest intersection: someone who can read a diff and a P&L. I build systems that reduce underwriting time by 30%, matching algorithms that cut deal-sharing time by 90%, and platforms that scale 10x. And I never forget the user on the other end.
When I'm not shipping, I'm on a Himalayan trail, running marathons, or building side projects like Notchd, a macOS app that puts your GitHub streak in the notch. The best PMs never stop building.
I use AI the way a good PM uses data: to move faster without cutting corners. At Recur Club, I've been using LLMs to accelerate everything from writing PRDs to reviewing underwriting logic, and the results show up in the work, not just the conversation.
Outside of work, every side project I ship now involves AI somewhere in the loop. Notchd was scoped, debugged, and shipped faster because I had a co-pilot that could context-switch as quickly as I could. Same with PopSlack.
The GitHub contribution graph below is my work account. Each green square is a day I pushed something. The cadence has picked up steadily as AI tools have gotten better, because the gap between thinking something and shipping it keeps shrinking.