About
Engineering leader, 20+ years in. Currently directing teams at Two Chairs, building AI products that have to actually work in HIPAA. Before that: scaling Edmodo from 1M to 50M, CTO-ing Backbone PLM from 3 to 30+ engineers, and a long tail of startups where I learned the hard way that the architecture diagram is not the system.
Right now
Running engineering at Two Chairs, building AI products in HIPAA. The flagship is AI Scribe — an LLM documentation tool that 75% of our clinicians actually use, and that saves them ~30 minutes per session note. Therapists liking an AI tool is rarer than you'd think.
The backstory
Wrote core infrastructure at Edmodo while the platform climbed from 1M to 50M users (turns out the things that work at 1M do not all work at 50M — we learned that several times). Then CTO at Backbone PLM, where I grew engineering from 3 to 30+. Consumer, enterprise, healthcare — same pattern: figure out what's worth building, then help the team actually build it.
How I lead
Engineers do their best work when they feel safe taking shots. Most of my job is removing the things that make that hard — unclear scopes, surprise meetings, silent disagreements. I've hired and grown engineers from junior to staff at four companies; the good ones still email me about things they're working on, which I take as a strong signal.
What gets me up
AI that has to work in production, in front of real users, in a regulated industry. The combination of constraints — clinical safety, latency, data governance, and a human on the other end whose afternoon you're either improving or ruining — is genuinely interesting. The fun is making it boringly reliable.
Education
B.S. in Computer Science & Political Science, University of San Francisco. Two majors because I couldn't pick.