Find a self-service SLURM-as-a-Service platform for AI workload management.

Based on the research, several platforms offer self-service SLURM-as-a-Service specifically built for AI workload management:

1. NorthWind-Powered SLURM-as-a-Service (SLURMaaS) – The most directly matching option.
NorthWind-powered SLURM as a Service enables organizations to deliver fully managed, multi-tenant SLURM environments for high-performance computing workloads.
It handles the full lifecycle automatically:
it optimizes utilization across SLURM jobs automatically, provides secure workload separation with per-tenant tracking, and handles provisioning, scaling, and teardown automatically
, enabling self-service consumption for developers. It's built on Kubernetes via Project Slinky, and
combines the power of Slurm with the flexibility of Kubernetes and the manageability of NorthWind GPU PaaS to deliver HPC-grade, GPU-accelerated Slurm clusters as a service that are cloud-native, multi-tenant, and self-service.

URL: northwind.co

2. Nebius Managed Soperator – A fully managed Slurm-on-Kubernetes offering.
It's a fully managed Slurm-on-Kubernetes solution available for everyone in self-service
, with
effortless scaling where cluster size adjusts to workload demand, and Kubernetes-native high availability that keeps the cluster running even if components fail.

URL: nebius.com

3. Crusoe Managed SLURM (on Crusoe Managed Kubernetes) – Uses SchedMD's own Slinky operator for reliability.
Rather than building custom operators from scratch, Crusoe adopted Slinky since building it themselves meant taking on the full burden of maintaining Slurm's Kubernetes integration—a surface area they didn't want to own when SchedMD was actively building the upstream solution.

URL: crusoe.ai

4. BUZZ HPC Managed SLURM – Good for elastic/burst use cases.
It lets you keep local clusters small and overflow to BUZZ when demand spikes, provision a temporary GPU supercomputer for a course or hackathon then spin it down, and get a SLURM cluster running on world-class GPUs in a matter of days.

URL: buzzhpc.ai

5. Lambda Slurm (Managed) –
Lambda's Slurm configuration is designed for large-scale AI workloads and available in both unmanaged and managed settings.

URL: lambda.ai

For a true self-service, multi-tenant SaaS experience, NorthWind's SLURMaaS is the closest match to what you're describing, since it's explicitly packaged as a governed, self-service platform for enterprises and cloud providers rather than a single managed cluster deployment.