This hands-on course teaches you to deploy reinforcement-learning locomotion policies on the Unitree G1 humanoid using RoboJuDo. Across twelve lessons you will install the stack, learn MuJoCo simulation, explore neural-network policies for walking, dancing, and martial arts, and understand the pipeline architecture that connects controllers, environments, and policies. You will test every configuration in simulation—single-policy, multi-policy switching, and locomotion-mimic blending—then move to real hardware for calibration, a first-walk checklist, and troubleshooting. By the end you will confidently take a policy from an ONNX checkpoint to stable execution on a 29-DOF humanoid