I talk about tech, real estate, finance, and immigration.

There are a great many real estate investors who have never built a model, and a great many data scientists who have never signed a lease. I sit in the narrow overlap, and it shapes everything about how I operate.

For nearly four years I led data science for the mobile experience at Realtor.com, running more than two hundred experiments on how Americans actually search for a place to live. That work gave me an unusually direct view of housing demand — not the version in the headlines, but the version in the clickstream.

I now build LLM and machine learning systems at Meta, after doing the same at Amazon on fulfillment network design, at Uber on a $100M fleet incentive program, and at Ola where I established the analytics function for one of India's largest ride-hailing markets.

Alongside all of it I've been buying and operating residential property since 2019 — long-term rentals in Ohio and Florida, a short-term rental in the Yosemite gateway, and residential holdings in India. Every acquisition goes through the same discipline I'd apply to a production model: price the risk deliberately, underwrite on evidence rather than narrative, and stress-test before you bid.

I partner with operators and allocators who value data, transparency, and a long time horizon.