Changyeon Kim

Changyeon Kim is a Ph.D. student at KAIST advised by Kimin Lee and Jinwoo Shin, and a research scientist intern at the NVIDIA Seattle Robotics Lab. His research centers on reinforcement learning for robotic agents, focusing on offline RL, reward learning, and long-horizon manipulation. He is the lead author of ACCRUE and DEAS, the latter accepted to ICLR 2026 for its work on scalable offline RL with vision-language-action models.