End-to-end implementation of Action Chunking Transformers (ACT) for imitation learning in robot manipulation tasks. Trained and evaluated in MuJoCo simulation on a pick-and-place task using the ROBOTIS FFW humanoid robot.