About

I started in English. A master's degree in literary theory and a license to teach high school taught me to read closely, argue from evidence, and explain hard things plainly. Those habits turned out to matter more in operations than anyone would guess.

In 2010 I joined Bonobos when it was doing under $10 million a year, and helped build the customer operations that carried it past $100 million. At Etsy I ran fraud and risk operations for more than $2 billion in annual payments, and built my first closed loop: fraud outcomes fed straight back into detection, so the system got sharper every week. After that came HavenLock, Summersalt, and Bodily, building fulfillment, supply chain, and data systems from scratch, usually while the business was scaling fast around them.

The pattern was always the same. Growth exposes every system that can't check its own work. So when AI arrived in operations, I saw the problem quickly: models that sound right, with nothing in place to prove they are.

In 2025 I went back to school on purpose, at Kellogg, Google, Microsoft, Michigan and Imperial, and founded Chorus AI Systems, my studio for AI that can't afford to be wrong. Every system I build there follows one rule: no claim reaches a person until it has been checked against the facts. CASDAM, BRAG, Dossier, and the Jev experiment all come from that idea. You'll see the Chorus label on them across this site.

Alongside the systems, I'm writing AI Operations Management, a textbook for running AI as the metered operating resource it is, and building the Applied AI Case Study Library to test those ideas against what has actually happened. I live in Chicago.