PeopleofSuperintelligence

Researcher

Yulong Cao

Yulong Cao is a computer scientist whose work tests the security of autonomous-driving perception and studies models that reason about road scenes. He earned B.S. degrees from the University of Michigan and Shanghai Jiao Tong University in 2017; his computer science and engineering Ph.D. at Michigan was advised by Morley Mao. His research spans LiDAR spoofing and removal attacks, adversarial trajectory prediction, and driving models such as Dolphins and Alpamayo-R1. As of October 2026, he is a research scientist in NVIDIA's ASPIRE Group, led by Marco Pavone, with interests in trustworthy AI, alignment, and embodied AI systems.

Now at
NVIDIA
Badges

Timeline

  1. 2017

    Michigan and Shanghai degrees

    In 2017, he earns B.S. degrees from the University of Michigan and Shanghai Jiao Tong University. At Michigan, his computer science and engineering Ph.D. is advised by Morley Mao.

  2. 2019

    A LiDAR spoofing attack

    In 2019, he leads the first security study of LiDAR-based perception in autonomous vehicles in the labs of Z. Morley Mao and Kevin Fu at the University of Michigan. The team spoofs obstacles in front of a victim vehicle with about 75% success; its paper, “Adversarial Sensor Attack on LiDAR-based Perception in Autonomous Driving,” is presented at the ACM Conference on Computer and Communications Security in November 2019.

  3. 2020

    LiDAR security and public research

    In March 2020, he and Mao write for Popular Science about the University of Michigan's RobustNet Research Group LiDAR work. In 2020, he, Jiachen Sun, Qi Alfred Chen and Z. Morley Mao present a USENIX Security paper showing that ignored occlusion patterns in LiDAR point clouds make self-driving cars vulnerable to spoofing. Their black-box attack reaches about 80% mean success, and the CARLO defense cuts the rate to 5.5%.

  4. 2022

    Adversarial trajectory prediction

    In 2022, he is first author of AdvDO, an ECCV paper on realistic adversarial attacks for trajectory prediction. In another 2022 paper, he is first author of Robust Trajectory Prediction against Adversarial Attacks, selected for oral presentation at CoRL with Marco Pavone among the coauthors.

  5. 2023

    Physical removal attacks

    In 2023, he coauthors “You Can't See Me: Physical Removal Attacks on LiDAR-based Autonomous Vehicles Driving Frameworks.” In moving-vehicle scenarios, the attacks remove 90% of a target obstacle's cloud points with a 92.7% success rate.

  6. 2023–2024

    Dolphins for conversational driving

    In December 2023, he coauthors Dolphins with Yingzi Ma, Jiachen Sun, Marco Pavone and Chaowei Xiao. The model builds on OpenFlamingo and adds a Grounded Chain of Thought process to serve as a conversational driving assistant; its paper appears at ECCV in 2024.

  7. October 2025–January 2026

    Alpamayo-R1 for autonomous driving

    In October 2025, he is one of 42 NVIDIA authors of Alpamayo-R1, a vision-language-action model that combines Chain of Causation reasoning with trajectory planning for complex driving scenarios. On-vehicle road tests report real-time performance at 99 ms latency, and the January 2026 revision lists him as an author.

  8. As of October 2026

    Research scientist in ASPIRE

    As of October 2026, he is a research scientist in NVIDIA's Autonomous Systems and Physical AI Research (ASPIRE) Group, led by Marco Pavone. His research interests include trustworthy AI, alignment and embodied AI systems.

Career

  1. NowNVIDIAResearch Scientist

Education

  1. 2017University of MichiganB.S.
  2. 2017Shanghai Jiao Tong UniversityB.S.

From the biography.

  1. 16% of AI career moves in the 2020s went from a university to a companyIncludes Yulong Cao. We estimate that 16% of AI career moves in the 2020s went from a university to a company; 29% of moves starting at a university went to a company, and 76% ended the university role.
  2. Google has the highest combined h-index among AI organizationsIncludes NVIDIA. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  3. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes NVIDIA. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.
  4. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes NVIDIA. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.

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 Yulong Cao yet. Send us a free-licensed photo.

  1. Yulong Cao | NVIDIA ASPIRE Group research.nvidia.com. Checked 22 October 2026.
  2. Adversarial Sensor Attack on LiDAR-based Perception in Autonomous Driving arxiv.org. Checked 22 October 2026.
  3. New attack on autonomous vehicle sensors creates fake obstacles wevolver.com. Checked 22 October 2026.
  4. Self-driving cars still have major perception problems | Popular Science popsci.com. Checked 22 October 2026.
  5. Towards Robust LiDAR-based Perception in Autonomous Driving: General Black-box Adversarial Sensor Attack and Countermeasures | USENIX usenix.org. Checked 22 October 2026.
  6. AdvDO: Realistic Adversarial Attacks for Trajectory Prediction | NVIDIA Research research.nvidia.com. Checked 22 October 2026.
  7. Robust Trajectory Prediction against Adversarial Attacks research.nvidia.com. Checked 22 October 2026.
  8. You Can't See Me: Physical Removal Attacks on LiDAR-based Autonomous Vehicles Driving Frameworks | USENIX usenix.org. Checked 22 October 2026.
  9. Dolphins: Multimodal Language Model for Driving arxiv.org. Checked 22 October 2026.
  10. ECCV Poster: Dolphins: Multimodal Language Model for Driving eccv2024.ecva.net. Checked 22 October 2026.
  11. Alpamayo-R1: Bridging Reasoning and Action Prediction for Generalizable Autonomous Driving in the Long Tail arxiv.org. Checked 22 October 2026.
  12. Yulong Cao yulongc.xyz. Checked 22 October 2026.