I currently am a Robotics & ML Software Engineer at NVIDIA where I prototype and develop ML-based motion planning and control strategies for robots of varied embodiments. I previously contributed as a Robotics Intern at NVIDIA (June–December 2024), designing parallelizable motion planning algorithms for manipulators in uncertain settings and leading the development of a modular benchmarking framework to evaluate universal manipulation planners.

As a Master's student at the University of Washington, I specialized in Robot Learning and Optimization. My research focused on combining Reinforcement Learning (RL) and Model-Based Optimization methods to develop general-purpose, reactive robot control systems. As a researcher in the UW Robot Learning Lab (UW CSE), led by Byron Boots, I have collaborated with Mohak Bhardwaj and Balakumar Sundaralingam on problems such as teaching robot manipulators to perform non-prehensile tasks with limited demonstrations and sparse rewards using the integration of Model-Predictive Control (MPC) and offline RL. Additionally, my work also includes unifying tactile sensing with robotic manipulation to address challenges like "object slip" and leveraging machine learning techniques for slip management. This research was conducted in the MACS Lab in the Mechanical Engineering Department under the guidance of Xu Chen.