In this session, we walk through deploying GPU-backed Kubernetes clusters with NVIDIA Run:ai on AWS using vCluster. You'll see how to set up the host cluster, provision isolated tenant environments with their own Kubernetes control planes, and pair Run:ai's GPU scheduling with vCluster's virtual clusters so each team gets full autonomy without fragmenting your infrastructure. We'll cover both hard and soft isolation models, show Private Nodes and Shared Nodes in action, and demonstrate how to keep GPU utilization high while giving platform teams the governance they need.

‍