Launch & Scale

Use these guides when you need to start runs beyond a single default machine or connect RLinf to physical hardware.

Guide	What you get
Multi-Node Training	Launch a run across multiple nodes.
Heterogeneous Clusters	Configure mixed hardware and node groups.
Cloud-Edge Collaboration	Split inference and training across cloud and edge.
Real-World Robots	Run RL on physical robot hardware.
Teleoperation	Let an operator take over from the policy mid-rollout.