Hi! I am a Ph.D. student in Mechanical Engineering at the University of California, Berkeley. I work in the ICON Lab under Prof. Negar Mehr. My research focuses on robot learning, especially on developing autonomous robots that can work alongside humans and other robots.

Prior to Berkeley, I earned my B.S. in Mechanical Engineering from Yonsei University. During my undergraduate studies, I worked as a research intern in MLCS Lab under Prof. Jongeun Choi, where I worked on MPCs and RL for autonomous driving, as well as visuamotor policy learning for contact-rich manipulation.

Research

My research lies at the intersection of robot learning, control, and foundation models. I study how LLMs and VLMs can guide reinforcement learning to acquire complex, coordinated robot behaviors. This includes building curriculum-driven frameworks such as CRAFT for multi-robot collaboration and extending them toward humanoid whole-body control and multi-humanoid interaction.

In parallel, I explore scalable imitation learning methods that can transfer single-robot expertise to multi-robot settings, where data collection is especially challenging. Ultimately, I aim to develop autonomous systems that can learn efficiently, coordinate intelligently, and operate alongside humans and other robots.

News

Publications

Preprint

CRAFT: Coaching Reinforcement Learning Autonomously using Foundation Models for Multi-Robot Coordination Tasks

S. Choi*, K. Ryu*, J. Ock, and N. Mehr (*Equal contribution)
Submitted to ICRA, 2026.

Arxiv Website

Preprint

MIMIC-D: Multi-modal Imitation for Multi-agent Coordination with Decentralized Diffusion Policies

D. Dong*, M. Bhatt*, S. Choi, and N. Mehr (*Equal contribution)
Submitted to ICRA, 2026.

Arxiv Code Website

Preprint

EquiContact: A Hierarchical SE(3) Vision-to-Force Equivariant Policy for Spatially Generalizable Contact-rich Tasks

J. Seo, A. Kruthiventy, S. Lee, M. Teng, X. Zhang, S. Choi, J. Choi, and R. Horowitz
Submitted to RA-L.

Arxiv Code Website