Scaling Physical AI & Robotics Systems with Ray

Build an end-to-end physical AI workflow using distributed systems powered by Ray.

This hands-on training session works through robotics data preparation at scale with LeRobot, MCAP, and MP4, Vision-Language-Action model fine-tuning with distributed training, simulation-based evaluation with Isaac Lab, agent-assisted pipeline development with the Anyscale physical AI skill, model distillation for on-edge real-time inference, and world-foundation model pre-training at scale, with the action running live on a real robotic arm.

You'll leave with working patterns for taking robotics pipelines from data preparation to deployment.