Reinforcement learning (RL) for robotics is often associated with large GPU clusters, distributed infrastructure, and x86-based development environments. Training a humanoid robot with high-fidelity ...