About

Most of what I build starts as a complaint.

Enterprise planning software costs a fortune, takes two quarters to change, and hands you a model you cannot inspect. Materialized views recompute a million rows because one row moved. A résumé gets ninety seconds of skimming and then a guess. Every one of those was something I lived with for years before I decided to go and write the alternative — most of it on the back of nineteen years spent at the seam between finance and engineering, the place where a planning model meets a data pipeline and somebody has to make both of them true at once.

So: planWell, an open-source EPM tool, built by someone who administered the commercial ones long enough to know which features are load-bearing and which are sales demos. duckDBSP and pg_dbsp, two attempts at the same idea — that a materialized view should cost you the delta and not the dataset — one inside DuckDB, one inside Postgres. Work With Me, a desktop AI coworker that runs on whichever model you choose, because the alternative is renting your tools from a single vendor. And a longer tail of things I built because I wanted them to exist: a study tracker that measures retention instead of coverage, an offline PDF editor, a pond full of neural networks that teach themselves to forage.

MBA in Entrepreneurship and Product Management from UC Berkeley Haas. B.Tech in Electrical and Electronics Engineering from NIT Calicut. Master Anaplanner — Anaplan’s highest credential — and a Certified ScrumMaster. I live in Concord, California. If something here is useful to you, or you think I am wrong about any of it, write to me.