I am a PhD candidate in Electrical Engineering at the Pennsylvania State University, working on learned policies and control for robotic systems. Most of my current work is on generative AI for robotics: vision-language-action models, diffusion and flow-matching policies, and reinforcement learning; and on making those policies dependable enough to run on hardware, which is where safety-critical and adaptive control come in.

I work across both sides of that boundary. On the learning side: policy-gradient and model-free reinforcement learning, imitation from demonstration, and large behaviour models conditioned on language and vision. On the control side: nonlinear and optimal control, adaptive and robust design, convex optimization, model predictive control, state estimation and motion planning. These carry over to most settings where a physical system has to act under uncertainty, and I am interested in those problems broadly, not only in the robots I have used to study them.

I build the software as well as the theory. I work daily in C++, Python, MATLAB and C# with ROS and PyTorch, and I test on real hardware, including quadrupeds, manipulators, mobile bases and full-size humanoids. My earlier research was on legged locomotion and humanoid gait control.