I am a third-year Ph.D. student in Information Science at Cornell University, advised by Saleh Kalantari, with committee members Kuan Fang and Tapomayukh Bhattacharjee. I design and evaluate embodied systems that perceive, interact, and learn from the physical world, from robot bimanual dexterous manipulation to data-automation pipelines, with applications in assistive and home robotics and healthcare. Previously, I received a Master of Architecture from UCLA and a B.S. in Architecture from Georgia Tech, and worked as a UX designer at Kuaishou Technology.
I am seeking Summer 2027 research internship opportunities in robot learning, dexterous manipulation, human-object interaction, and embodied AI.
Preprint 2026
Pre-trains a bimanual dexterous policy on 1M human-video clips by reconstructing scene geometry alongside hand motion, then adapts it to a robot with a few demonstrations.
ECCV 2026
A multimodal egocentric navigation framework that jointly forecasts human trajectories, head motion and cognitive states from first-person video and multimodal sensory cues.
In progress
Generates full-body human interaction with articulated objects from language, predicting the intended state change, enforcing it during generation, and verifying the result.
In progress
DexTele is a VR-based platform for dexterous teleoperation and robot data collection. We are developing physics-informed contact retargeting that maps human hand contacts onto the robot hand, to collect high-fidelity demonstrations for dexterous robot learning.
CHI 2026
Evaluates the trade-offs in spatial learning, wayfinding performance, cybersickness, and presence of controller-based VR locomotion techniques in virtual indoor environments.
npj Digital Medicine 2025
A review of 56 studies finds LLMs are used in oncology to summarize and communicate cancer information, but need better clinical readiness, safety evaluation, and datasets for standardized assessment.
Advanced Engineering Informatics 2024
An AR app that improves older adults' wayfinding efficiency and spatial learning when navigating complex indoor environments.
Selected UX and architecture work.
