Haohong Lin

Haohong Lin

PhD Candidate

Carnegie Mellon University

Biography

I’m an ML researcher and practitioner at Tesla AI, working on the FSD. Prior to my full-time career at Tesla, I was a PhD student at SafeAI Lab and a Master in Machine Learning at School of Computer Science of Carnegie Mellon University, advised by Professor Ding Zhao. I also worked closely with Professor Bo Li from UIUC. My research interest lies in the causality-driven approaches for safe and generalizable Physical AI systems, with some exciting applications in autonomous driving and embodied agents. During my PhD, I’ve spent my summers/falls as research intern at NVIDIA, Waymo, Cruise, and MERL. Prior to CMU, I receive my B.S. Degree with highest honor in Automation (Robotics Track) from Zhejiang University in 2021.

Education

  • PhD at SafeAI Lab, 2021-2026

    College of Engineering, CMU

  • MS in Machine Learning, 2023-2024

    School of Computer Science, CMU

  • BS in Automation (Robotics), 2017-2021

    Department of Control Science and Engineering, ZJU

Recent News

All news»

[09/2025] One paper about policy adaptation for end-to-end driving accepted by NeurIPS 2025, see you in San Diego!

[08/2025] Starting my part-time intern at NVIDIA Research autonomous vehicle research group.

[05/2025] Starting my summer intern at AI research team of Waymo LLC!

[03/2025] I have been awarded the William and Alice McGaw Graduate Fellowship at CMU!

[02/2025] One paper about causal reasoning and diffusion-based scenario generation is accepted by CVPR 2025!

Recent Publications

Quickly discover relevant content by filtering publications.

Pushing Forward Pareto Frontiers of Proactive Agents with Behavioral Agentic Optimization

ICML 2026
Pushing Forward Pareto Frontiers of Proactive Agents with Behavioral Agentic Optimization

WestWorld: A Knowledge-Encoded Scalable Trajectory World Model for Diverse Robotics

ICRA 2026
WestWorld: A Knowledge-Encoded Scalable Trajectory World Model for Diverse Robotics

CrashAgent: Crash Scenario Generation via Multi-modal Reasoning

Under Review
CrashAgent: Crash Scenario Generation via Multi-modal Reasoning

Contact

  • 5000 Forbes Avenue, Pittsburgh, PA 15123
  • Hamerschleg Hall, Carnegie Mellon University
  • My LinkedIn