Things I've built that solve real problems.
Built a student attendance tool during my final year of college. 57,000 users later, I still haven't spent a rupee promoting it.
Research and writing on trade, AI governance, economics, and whatever else I'm curious about.
Automated the entire meeting workflow for a 100-franchise education company. Scheduling, approvals, attendance, certification, all coordinated through a chatbot.
I published a book on practical AI in September 2023. Somehow, it ended up on a Union Minister's desk.
Presented to Union Minister Piyush Goyal (Commerce & Industry)
Two friends preparing for CA Finals, one dashboard, and a simple rule: hit start when you study. Live hours. Live status. If you're idle and your friend is working, everyone knows. 25 weeks, 1,239 hours.
Specialization is the greatest invention of civilization. It's also its most dangerous blind spot.
The accountant sees cash flows. The engineer sees architecture. The policymaker sees incentives. Each of them is right. None of them is complete. And the problems that matter most almost always live in the spaces between them.
This isn't a new observation. Adam Smith saw it in 1776: the same division of labor that multiplied productivity by orders of magnitude also made individual workers, in his words, "as stupid and ignorant as it is possible for a human creature to become." Expertise narrows. That's not a moral failing. It's a structural one. Your training teaches you what to see, and in doing so, it teaches you what to ignore.
The more I read, the more I notice the same structures showing up in places that aren't supposed to be related. Rousseau asked where legitimate authority comes from. Two and a half centuries later, when algorithms decide who gets credit and what content gets seen, we are asking his exact question. Someone working on AI governance who hasn't grappled with Rousseau is solving a problem they haven't fully understood. Someone who has read Rousseau but doesn't understand how a language model works can't see where the old answers break.
I believe one of the most valuable skills for the coming decades will be translation: the ability to move between specialized worlds fluently enough to see what each one is missing. Not knowing everything. Recognizing that the same structures, feedback loops, incentive gradients, information asymmetries, keep appearing in different costumes, whether in a balance sheet, a codebase, or a trade agreement.
I'm still learning how to do this. But it's the direction everything I've done so far seems to point.
Most interesting problems live at intersections.
Always happy to talk.