Seohong Park
Hey! I'm a Ph.D. student at UC Berkeley advised by Sergey Levine. I'm also currently working at Physical Intelligence. I'm interested in developing scalable reinforcement learning (RL) algorithms. More specifically, my goal is to make off-policy RL scalable so that we can apply RL to many more real-world problems, such as robotics. To this end, I've been mainly thinking about data-driven RL and pre-training for RL. CV | Email | Twitter | Google Scholar | Github Is RL Scalable? @ Guest Lecture for CS 185/285, UC Berkeley (April 2026) Is RL Scalable? @ Guest Lecture, Georgia Tech (April 2026) Is RL Scalable? @ Amazon FAR (Jan 2026) Scalable Data-Driven RL @ AIRe Lab, Carnegie Mellon University (Mar 2025) Scalable Data-Driven RL @ RL Lab, Princeton University (Mar 2025) Scalable Data-Driven RL @ REAL Lab, Stanford University (Mar 2025) Toward Scalable Unsupervised RL @ Intel AI Lab (Feb 2024) Unsupervised Reinforcement Learning @ Reading group on Unsupervised RL, MILA (Apr 2023) Ph.D. stu