Harsh Gupta United States flag

Last updated: August 2026

I'm a graduate student at Stanford, pursuing an M.S. in Computer Science. I work with Prof. Shuran Song at REAL on dexterous manipulation and world models.

Previously I did my B.S. in Computer Engineering at UIUC, where I worked with Prof. Wenzhen Yuan and Prof. Guanya Shi (at CMU) across tactile sensing, cross-embodied learning, and robot manipulation.

Harsh Gupta profile photo

/ Resume / Scholar / LinkedIn / Github / Twitter

Selected Research

I’m interested in learning manipulation from large-scale human video (YouTube) for the humanoid form factor (dexterous hands). See full list of publications here.

LUCID: Learning Embodiment-Agnostic Intent Models from Unstructured Human Videos for Scalable Dexterous Robot Skill Acquisition
Harsh Gupta, Guanya Shi†, Wenzhen Yuan†
Under review at Conference on Robot Learning (CoRL), 2026
Paper / Webpage
UMI-on-Air: Embodiment-Aware Guidance for Embodiment-Agnostic Visuomotor Policies
Harsh Gupta, Xiaofeng Guo, Huy Ha, Chuer Pan, Muqing Cao, Dongjae Lee, Sebastian Scherer, Shuran Song, Guanya Shi
International Conference on Robotics and Automation (ICRA), 2026
Paper / Webpage / Code

† denotes equal advising.