I am currently studying Computer Engineering at the University of Toronto, with a focus on research at the frontier of physical AI. I am especially interested in combining world models and reinforcement learning to develop robust, generalizable policies for robots operating in the real world.

Previously, I worked at Berkeley Street Capital, where I developed and evaluated reinforcement learning agents for high-frequency quantitative trading and worked on improving market simulation for RL training. I also conducted research at STARS on human-guided reinforcement learning for dexterous manipulation, exploring how interventions and learned value functions can improve robotic policies.

Outside of research and engineering, I enjoy running, playing sports, and photography.

Email: kartik [at] mail [dot] utoronto [dot] ca