About

Hi there, I’m Zubin. I love building products at the intersection of technology, government, privacy and public safety. I’ve built over a hundred of them — for government agencies, Fortune 500s, and startups of my own — and I’m grateful the work has been recognized by Apple, Gartner, TechCrunch, the Wall Street Journal, ABC News, Inc., Reddit, and Entrepreneur.

Since 2020 I’ve been on the founding team at Docugami, teaching machines to actually read business documents — vision models, language models, and document graphs, with every answer citable back to the paragraph, table, or cell it came from. Before that I spent eighteen years at ImageWork Technologies, where I ran solutions engineering and architected ClaimConnect, a claims platform that moves $2.1 billion a year for New York City’s police officers and firefighters.

In between, the ventures: CiviGuard (emergency communication; NYC’s Advance Warning System), Quipio (typographic art with a send button; featured twice by Apple), Alma Safety (911 location accuracy — our findings influenced FCC requirements), and now Barkie and Murch, both applying multimodal AI to things I genuinely care about: golf, and collections.

I took an MBA at MIT Sloan as a Sloan Fellow (2020), hold six US patents in document intelligence, wrote Pro JSF and HTML5 for Apress, and contributed to O’Reilly’s 97 Things Every Software Architect Should Know — there’s a 2009 post here about that one.

Away from keyboards: ultra-marathon open-water swimming (the English Channel) and a black belt in Budokan karate.

About the name. Any entangled pair of particles is correlated, but a maximally entangled pair — a Bell pair, like the |ψ⟩ = (|00⟩ + |11⟩)/√2 in this site’s footer — is a stranger thing. Examine either particle on its own and you find pure noise: a fair coin, no structure at all. Every bit of information lives in the correlation; measure one and the other’s outcome is fixed, however far apart the two have drifted.

Reach me: zwadia@gmail.com · linkedin.com/in/zubinwadia