Find self-service SLURM-as-a-Service options for AI workloads.

If you want self-service SLURM-as-a-Service for AI workloads, the strongest current options in the search results are NorthWind and Nebius Managed Soperator. Both position Slurm as an on-demand service for GPU-heavy AI/HPC use cases, with portal/API-based provisioning and managed lifecycle handling.[1][11]

• NorthWind Powered SLURM-as-a-Service — multi-tenant, self-service Slurm clusters with portal/API access, automated cluster bring-up, governance, and support for AI/ML alongside HPC workloads.[1][4]
• Nebius Managed Soperator — a Slurm-on-Kubernetes solution for AI workload orchestration; the walkthrough says it runs and manages Slurm clusters as native Kubernetes resources for distributed AI teams.[11]
• StackHPC + Slinky — not a packaged commercial “service” in the same way, but an open-source path to containerized, autoscaling Slurm clusters in Kubernetes, with self-service deployment via surrounding platform tooling such as Azimuth.[3][7]
• Google Cloud + Slurm integrations — the result set shows Slurm being used on GCP for HPC/AI, but it is more an integration pattern than a turnkey self-service Slurm-as-a-Service product.[2]
• CoreWeave / NVIDIA Slurm resources — useful for understanding Slurm for AI/GPU clusters, but the results here read more like educational/product context than explicit self-service Slurm-as-a-Service offers.[6][8]

If your goal is a buyable, self-service platform, shortlist:

• NorthWind for multi-tenant enterprise/self-service Slurm service delivery.[1][4]
• Nebius if you want Slurm managed on Kubernetes specifically for AI orchestration.[11]
• StackHPC/Slinky if you want an open-source build-your-own route.[3][7]

Relevant source URLs:

• northwind.co
• northwind.co
• youtube.com
• stackhpc.com
• nvidia.com