Uber TL for 40 enginers building the next generation of agentic products at Airtable.

Previously TL for field agents, took product 0-1. Built most of our AI async serving architecture like semantic search.

Worked with the Advanced Materials and Discovery division and partner labs at Stanford, MIT, SLAC, and others on some fun projects:

• Taking battery test times from months to days via AI predictions, automated test writing, and more

• Higher throughput polymer synthesis through predictive modeling and easier data exploration

• Other stuff I can't talk about

Learned a lot about batteries, writing compilers, operating data pipelines, how to run a good user study, creating useful and performant data visualization tools for large (tB) datasets

Startup I worked on with a good friend trying to crack automating most of 1-1 math tutoring (but not teaching, the former is easier). Based on field tests we think we cracked the basic student-tutor loop but found we needed more years of productization work than we had runway. I hope to come back to this someday if the idea doesn't get scooped. Had a really neat custom computer algebra system we wrote that heavily abused v8 to keep performance real time (turns out kids also hate slow UIs).

• Designed and prototyped low cost solution to remote logging of environmental data

• Programmed device to interface with 3rd party sensors and sync

• Created dashboard to monitor incoming sensors and view incoming data

My life before software: ran a tutoring business, after school programs, wrote curriculum, summer camps. Lots of fun stuff.