About

I'm a product manager building AI products at the intersection of trust and creativity. At Adobe, I work on making Content Credentials enterprise ready: figuring out how to productize and monetize content provenance for the large creative teams that make the world's content.

The interesting part is finding use cases worth paying for. One I like: tracing a JPEG back to the Photoshop file it came from, so a team can navigate its content supply chain and see everywhere an asset has been used. I build proofs of concept with enterprise partners across 10+ Creative Cloud teams, then work with sales to turn them into a generally available product that holds up across industries. There's more on the underlying technology at contentcredentials.org.

How I got here

I grew up in Hyderabad, and came to the US for school. I studied computer science at the University of Wisconsin–Madison, by way of a couple of years at the University of Washington. Somewhere in there I worked as campus tech support, did a stint as a data scientist studying what Pokémon GO was doing to people, and built an app with an interactive campus map that found the shortest walking route between buildings. I have apparently been interested in walking routes for a long time.

I started my career at UnitedHealth Group, where I spent four years moving from backend software engineer to technical product manager. I led an AI/ML personalization platform serving 13.4 million Medicare members, managed a team of 15 engineers, and drove a 300% increase in engagement. I moved into product from inside the engineering team, which is still the most useful thing about how I work.

I went to Duke's Fuqua School of Business for an MBA, graduating in 2025. Along the way I led Tech Club recruiting, was lead TA for managerial economics, and co-led the improv club, which earned me a Paired Principle Award. I spent a summer at Adobe as an AI product management intern working on Acrobat's AI Assistant, and came back full-time.

How I work

The engineering background is the thing that shows up most. I'm hard to bluff on feasibility, I write my own SQL, and I tend to argue for the boring instrumentation work that makes every later decision cheaper: the debugging tool nobody asked for, the eval dataset that turns a model argument into a measurable one.

I also think a lot about what a metric is actually claiming. My side project measures park coverage against the path network instead of an area grid, because the easy version of that number would have been flattering and wrong. That instinct is more or less the whole job.

Outside of that

I watch Formula One and cricket with more emotional investment than is strictly reasonable, ski when I can, and play a rotating amount of badminton and pickleball. I also do improv, which has turned out to be startlingly useful training for product work, since both are mostly about listening carefully and committing to a bad idea fast enough to find the good one.

I run mock interviews and edit résumés for people trying to break into PM, tutor math, and once helped package fifty thousand meals in two hours. I'm always up for talking to someone earlier in that path than me.