Imitation learning project for robotic manipulation. The goal is to train policies in simulation, then deploy them on a real robot arm.