Summary
Pranav Agarwal is an applied scientist with a PhD in AI and nine years of experience specializing in reinforcement learning, world models, and robotics applied to autonomous driving and character animation. His research blends transformer-based world models, automated reward modeling from large offline trajectories, and continual learning to achieve sample-efficient RL for complex robotic systems. He has published and developed RL solutions for end-to-end driving in CARLA, worked on ego-vision captioning metrics at Inria, and built simulation-based crane automation during an MITACS fellowship. Now at Wayve, he focuses on bringing research-grade world models into real-world driving stacks, pairing academic rigor from Mila with hands-on production research internships. Notably, he combines deep learning and interpretability work—investigating decision-making in LLMs—so he can translate opaque model behavior into actionable insights for robotic systems.
9 years of coding experience
2 years of employment as a software developer
Doctor of Philosophy - PhD Artificial Intelligence, Doctor of Philosophy - PhD Artificial Intelligence at École de technologie supérieure
Senior Secondary, Senior Secondary at Creane Memorial High School
Bachelor of Technology (B.Tech.) Electrical Electronics and Communications Engineering, Bachelor of Technology (B.Tech.) Electrical Electronics and Communications Engineering at Indian Institute of Information Technology Guwahati
English