A record of taking systems from zero to scale.

The open web is the largest live record of the labor market there is — and none of it arrives as data. This engine does the whole read in one system: crawl the frontier, extract structure from raw pages, enrich the records until they hold together as labor-market data. Built toward reading 500M+ domains; I designed it end to end.

An AI agent forgets everything the moment its session ends, and a team that forgets cannot compound. This system is the memory that outlives the session — persistent memory, tickets, and files in one store, backing human and agent workflows alike. Most days my engineering team is a fleet of AI agents; this is what lets them work as a team.

The AI job-search product the two systems above exist to feed. Everything else on this page at Metaintro is machinery — this is the surface a job seeker actually touches, and where the machinery has to prove itself. It went from a thousand users to a million in four months.

Physical-climate risk analytics for real-estate portfolios: quantifying how exposed a portfolio actually is, in terms institutional investors and lenders can act on. I led engineering on it — the discipline of getting climate risk solid enough to sit under institutional decisions.

A commercial-real-estate platform: deal management, document workflows, and portfolio analytics for institutional clients. Commercial real estate runs on documents, and a deal moves only as fast as they do — the work was making deals, their documents, and the analytics on top of them behave as one system.

A personal AI assistant for my family — live across SMS, WhatsApp, voice, web, and push. One assistant that reaches each person on the channel they already use. Built and run solo.