PeopleofSuperintelligence

Researcher

Hang Zhao

Also known as 赵行

Hang Zhao is a researcher in autonomous driving, multimodal learning and embodied intelligence. He earned a B.S. at Zhejiang University in 2013 and a Ph.D. at MIT in 2019 under Antonio Torralba, then worked as a research scientist at Waymo from 2019 to 2020. His team's DriveVLM, a fast-slow dual driving system built on vision-language models, was deployed by Li Auto in hundreds of thousands of electric vehicles, and in 2023 he co-founded the humanoid robotics startup Galaxea AI. MIT Technology Review named him to its Innovators Under 35 list for China in 2024, and in 2025 he coauthored Galaxea's G0 dual-system model for robots. He is an assistant professor at Tsinghua University's Institute for Interdisciplinary Information Sciences and leads its MARS Lab.

Now at
Galaxea AI and Institute for Interdisciplinary Information Sciences
Badges

Timeline

  1. 2013–2019

    An MIT PhD under Torralba

    He receives a B.S. from Zhejiang University in 2013 and a Ph.D. from MIT in 2019 under Professor Antonio Torralba. In 2018, he is first author of The Sound of Pixels, which introduces PixelPlayer, a system that learns from unlabeled videos to locate the image regions that produce sounds and to separate the sound from each pixel.

  2. 2019–2020

    Research scientist at Waymo

    He works as a research scientist at Waymo, known as Google's self-driving project, from 2019 to 2020. In May 2020, Waymo describes VectorNet, a behavior-prediction model he develops with Jiyang Gao, Yi Shen and Chen Sun that represents road features and the paths of other road users as sets of vectors and needs less compute than convolutional networks.

  3. 2021

    The HDMapNet framework

    In 2021, he coauthors HDMapNet with Qi Li, Yue Wang and Yilun Wang. The framework learns HD semantic maps from onboard camera and LiDAR data by predicting vectorized map elements in the bird's-eye view, and it earns a best paper award at a CVPR 2021 workshop.

  4. 2023

    Galaxea AI and robot parkour

    In 2023, he co-founds the humanoid robotics startup Galaxea AI, where he serves as Chief Scientist. His paper Robot Parkour Learning, written with Ziwen Zhuang, Zipeng Fu, Chelsea Finn and others, trains one vision-based policy that lets low-cost robots climb, leap, crawl and squeeze through obstacles, and it is a Best System Paper finalist at CoRL 2023.

  5. 2024

    DriveVLM in Li Auto cars

    He coauthors DriveVLM, a driving system that uses vision-language models for scene understanding and planning, and DriveVLM-Dual, a fast-slow hybrid with the traditional driving pipeline; the paper appears at CoRL 2024. His industry partner Li Auto deploys the solution in hundreds of thousands of its electric vehicles.

  6. 2024

    Innovators Under 35 for China

    MIT Technology Review names him one of its Innovators Under 35 for China in 2024, citing the first foundation model-based fast-slow dual system for self-driving.

  7. August 2025

    The Galaxea G0 model

    In August 2025, he coauthors the Galaxea Open-World Dataset and G0 with Jiyang Gao, Huazhe Xu and colleagues. G0 is a dual-system framework that pairs a vision-language model for multimodal planning with a vision-language-action model for fine-grained execution.

  8. As of October 2026

    Professor and Galaxea co-founder

    After Waymo he returns to China, and he is an assistant professor at the Institute for Interdisciplinary Information Sciences (IIIS) at Tsinghua University and principal investigator of the MARS Lab. He remains a co-founder of Galaxea AI.

Career

  1. Since 2023Galaxea AICo-founder, Chief Scientist
  2. NowTsinghua UniversityAssistant Professor
  3. NowInstitute for Interdisciplinary Information SciencesPrincipal Investigator
  4. 2019–2020WaymoResearch Scientist

Education

  1. 2019Massachusetts Institute of TechnologyPh.D.
  2. 2013Zhejiang UniversityB.S.

From the biography.

  1. 11% of AI doctorate holders from 2010 to 2016 founded or led a company or lab by 10 years after graduationIncludes Hang Zhao. We estimate how often AI doctorate holders founded or led a company or lab by 10 years after graduation, how waiting times compare between founders and top jobs, and which universities trained the most.
  2. 23% of people shaping AI who left a university for a company later returned to a universityIncludes Hang Zhao. We estimate that 468 of the 2,011 people shaping AI who left a university for a company (23%) later returned to a university after a median of 5 years, and 163 of 829 (20%) made the reverse round trip.
  3. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes Tsinghua University. We estimate where the people shaping AI studied and where their careers led. Massachusetts Institute of Technology leads with 467 former students, including 291 recorded doctorates.
  4. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Tsinghua University. The median h-index is 48 at universities and 14 at frontier AI labs. We compare publishing careers, recent citations, dual appointments and job moves across 2,420 university researchers and 308 frontier-lab researchers.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Hang Zhao yet. Send us a free-licensed photo.

  1. Hang Zhao hangzhaomit.github.io. Checked 21 October 2026.
  2. [1804.03160] The Sound of Pixels arxiv.org. Checked 21 October 2026.
  3. VectorNet: Predicting behavior to help the Waymo Driver make better decisions waymo.com. Checked 21 October 2026.
  4. [2107.06307] HDMapNet: An Online HD Map Construction and Evaluation Framework arxiv.org. Checked 21 October 2026.
  5. Hang Zhao | Innovators Under 35 innovatorsunder35.com. Checked 21 October 2026.
  6. [2309.05665] Robot Parkour Learning arxiv.org. Checked 21 October 2026.
  7. 赵行-清华大学交叉信息研究院 iiis.tsinghua.edu.cn. Checked 21 October 2026.
  8. [2402.12289] DriveVLM: The Convergence of Autonomous Driving and Large Vision-Language Models arxiv.org. Checked 21 October 2026.
  9. [2509.00576] Galaxea Open-World Dataset and G0 Dual-System VLA Model arxiv.org. Checked 21 October 2026.