The one doing
all the asking.

I grew up moving between places, which is probably why I am comfortable walking into a room where nobody has written anything down yet. You learn to work out how a thing runs by watching it, asking a lot of questions, and being willing to look stupid for the first week.
I started in mechanical engineering, spent a stint in investment banking research, and then found product. None of it was a plan. What connects it is that I keep moving toward problems where the technical constraint and the human problem are the same conversation.
These days that means AI reliability: systems where being confidently wrong is expensive, and where the product job is making the uncertainty legible instead of hiding it. On the side I build small things, write, dance badly on the internet, and cook far more than I should.
Tamil Nadu to
Pittsburgh.
Carnegie Mellon University
M.S. Product Management
Tepper School of Business and the School of Computer Science, on a course run jointly by both. Currently at the Applied Generative AI Lab, building a scanner that checks whether a model's answer is actually supported before anyone reads it.
- OrthoBerryProduct Management InternRebuilt treatment planning end to end, and cut plan creation from 20 to 25 minutes down to about 3 in testing.
- Applied Generative AI LabLLM misinformation scannerBenchmarked factuality pipelines, then productised the most efficient one as a deployable guardrail.
Run along the line, or use the arrow keys. 1 of 10
Along the way








- Dance
- Led a 16-member team to second place at NITTFEST Choreonite. A lockdown YouTube channel, Melthalam, crossed 75,000 views.
- Writing
- Ink & Feather, plus a piece on buttons and prompts published in Bootcamp.
- Photography
- A Behance portfolio that gets updated in bursts, usually after a trip.
- Chess, anime, cooking
- In roughly that order of seriousness, though cooking is catching up.

