Architected an on-vehicle trigger evaluation system and led its development from conception to production. Reduced time to high-resolution vehicle data from 6 weeks to 15 minutes, tightening Tesla's firmware release cycle and creating a step change in vehicle data collection capabilities. Led a small team of engineers directly and drove a highly complex rollout to the entire global fleet (4M+ vehicles).

Led telemetry development for a skunk works project racing the Plaid Model S prototype at the Nürburgring racetrack in Germany. Using trackside learnings, built a novel telemetry system to live stream embedded systems data from vehicles. Architected the system to stream full-rate data, adapt to low bandwidth environments, and provide sub-second latency.

Designed distributed systems to transform, store, and analyze massive quantities of prototype vehicle signal data. Improved query times 2 orders of magnitude, decreased analysis error rates from 32% to < 1%.

Developed a search and analytics engine augmented with natural language understanding for a widely-used medical device database.