I lead the vision team at Autopilot! We're the brains behind Autopilot's vision only perception stack.
The team is a mix of software engineer generalists and deep learning specialists who are responsible for the development of neural networks for object detection + kinematics estimation, lane geometry prediction, generic depth estimation, traffic controls, etc. We train/quantize/deploy neural networks for our custom AI accelerator, improve inference latency via architecture search, autolabel fleet data, build evaluation + visualization tools, and much more!
We're always hiring strong engineers and AI scientists. Contact me if you're interested in joining our mission to achieve full self driving!
I get excited about things at the intersection of computer systems and computer vision, autonomous driving is the perfect fusion! Improving the accuracy and inference latency of our large multi-camera + video models with clever neural network architectures is totally my jam. Coaxing magic out of our AI accelerator is my version of Sudoku.
I was born and raised in the Midwestern USA, but transplanted to the Bay Area for a PhD in computer science at Stanford. I dropped out just two years later (one of my more reckless decisions) to join the then nascent Autopilot AI+Vision team with Andrej Karpathy. I have had a hand in just about every neural network that has gone to production in recent history and I keep myself closely connected to the technology.
I also run long distances for fun :). Look for me running along The Embarcadero in the mornings.