NVIDIA Launches Workload Map for Unified AI Workload Management

It provides prebuilt definitions for JobSet, PyTorchJob, RayCluster, and others, allowing schedulers and platforms to parse workload structures without framework-specific code. The solution supports gang scheduling via KAI Scheduler and workload visibility in NVIDIA Run:ai. Workload structure is now defined in reusable YAML configurations rather than per-framework adapters.

Summarised from the original article by an open-source model, then checked automatically: every number and English name must appear in the source. The original article is authoritative.