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

Here are the most relevant self-service SLURM-as-a-Service options for AI workloads I found:

• NorthWind SLURM-as-a-Service — Fully managed, multi-tenant SLURM with self-service cluster launch via portal or API, built for HPC and AI/ML workloads.[1][2] It emphasizes per-tenant isolation, governance, and GPU-accelerated Slurm clusters on Kubernetes.[1][2]
• Nebius Managed Soperator — A fully managed Slurm-on-Kubernetes solution available in self-service; Nebius says Soperator can run in Nebius AI Cloud or other Kubernetes environments and supports ML/HPC workloads.[3][16]
• Google Cloud Cluster Director for Slurm on GKE — A managed, opinionated Slurm experience for AI/HPC on GKE; Google also offers the Slurm Operator add-on for GKE for teams building custom AI/ML platforms.[12]
• Lambda Managed Slurm — Managed or unmanaged Slurm on 1-Click Clusters with NVIDIA HGX B200/H100 GPUs, positioned for AI workloads and “optimized for AI workloads.”[7]
• StackHPC SLURM-as-a-Service — More of a platform pattern / service offering than a packaged cloud product, but StackHPC describes SLURM-as-a-Service and self-service portal concepts for scientific and AI/HPC environments.[4][15]

Quick recommendation:

• If you want the strongest enterprise self-service + multi-tenant story, start with NorthWind.[1][2]
• If you want a managed cloud platform with Slurm already integrated, look at Nebius or Google Cloud.[3][12][16]
• If you want GPU cluster procurement plus managed scheduler, check Lambda.[7]

Source URLs:

• northwind.co
• northwind.co
• docs.nebius.com
• docs.cloud.google.com
• lambda.ai