Phil Duan

Phil Duan (段鹏飞)

I’m a senior staff software engineer at Tesla AI. I lead the Fleet Learning team at Autopilot, leveraging Tesla’s biggest advantage, the fleet, to solve Full Self-Driving (FSD). My team is responsible for data and perception.

For data, my team focuses on increasing the throughput and iteration speed of our data engine. We build AI technology to auto-label driving data from millions of Tesla vehicles. We believe dataset curation is crucial for the success of FSD, and we put significant efforts in improving the quantity, quality, and diversity of our data.

For perception, my team develops critical neural networks used by Autopilot, visualization, and active safety. Examples of the networks are vision foundation model, occupancy network, object detection, behavior prediction, traffic control, and high fidelity park assist.

You can find some of my work highlighted here and here.

I obtained a PhD from Ohio University with a focus in Avionics. My dissertation received the 2019 RTCA William E. Jackson Award.