Hang Zhao

Hang Zhao is dedicated to advancing embodied intelligence, with a focus on autonomous driving and humanoid robots. He previously worked as a scientist at Waymo, Google’s self-driving car company. Later he moved back to China and joined Tsinghua University as an Assistant Professor. 

Zhao sought to address the slow progress of autonomous driving in passenger vehicles, and proposed the vision-centric autonomous driving open-source framework VCAD. To better understand complex and long-tail driving scenarios (such as abnormal road conditions and pedestrian gestures), Zhao’s team proposed DriveVLM, leveraging the chain-of-thought reasoning capabilities of vision-language large models to understand complex driving scenarios and make driving decisions for autonomous vehicles. This was the first time that few-shot and zero-shot driving scenario understanding was achieved, significantly improving the autonomous driving system’s ability to handle long-tail scenarios. Furthermore, Zhao first proposed a practical large model-based autonomous driving solution: the fast-slow dual autonomous driving system DriveVLM-Dual. This system significantly enhanced the large model's spatial understanding capabilities and break through the large model’s reasoning speed bottleneck. His industry partner Li Auto, has deployed the solution into hundreds of thousands of their electric vehicles, marking the first mass-produced large models in autonomous driving. This is a rare “first proposed by China, first deployed in China” achievement, significantly accelerating the adoption of autonomous driving in consumer vehicles.

In 2023, Zhao co-founded the humanoid robotics startup GalaXea AI, where he serves as Chief Scientist.

On the way of seeking scaling law of embodied intelligence, Zhao drives breakthroughs in foundation models with the ultimate goal of bringing L4-level humanoid robots into everyday life.