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

Here are several solid self‑service SLURM-as-a‑Service platforms tailored for AI and HPC workload management:

• NorthWind-powered SLURM‑as‑a‑Service
Offers fully managed, multi-tenant SLURM clusters with a cloud-like, on‑demand experience. Tenants can launch SLURM clusters via portal or API, with built-in isolation and governance. (northwind.co)

• Crusoe Managed Slurm (on Crusoe Cloud)
Delivers managed SLURM orchestration atop GPU-optimized infrastructure via UI or CLI. Includes provisioning of Slurm control plane, login nodes, shared storage, node health checking, and automatic hardware remediation. (docs.crusoecloud.com)

• Google Cloud Cluster Director
A managed infrastructure service for deploying and overseeing SLURM (and Kubernetes) clusters on Google Cloud. Users manage resources via control plane, API, or CLI and only pay for underlying compute, storage, and networking. (cloud.google.com)

• AWS Parallel Computing Service (AWS PCS)
A managed HPC service that simplifies running Slurm-based AI/HPC workloads on AWS. Supports elastic, full-stack environments with managed updates and observability, reducing infrastructure overhead. (aws.amazon.com)

• Lambda Managed Slurm (1‑Click Clusters)
Offers AI-optimized Slurm clusters (with NVIDIA GPUs) in both managed and unmanaged modes. The managed variant includes patching, monitoring, job history tracking, container support, and SchedMD-backed technical support. (lambda.ai)

Optional mention: Waldur (open-source) integrates SLURM with OpenStack, offering a user self‑service portal for HPC and cloud environments, along with billing and support desk features. (waldur.com)

Summary Table:

| Platform | Model/Environment | Highlights |
|-------------------------------|-----------------------------|----------------------------------------------------------|
| NorthWind SLURM‑as‑a‑Service | Multi‑tenant cloud‑like | Self‑service portal/API, automated provisioning |
| Crusoe Managed Slurm | GPU‑optimized private Cloud | Full SLURM stack via UI/CLI, hardware failure handling |
| Google Cloud Cluster Director | Public cloud (GCP) | Managed SLURM + Kubernetes, pay-per-resource |
| AWS Parallel Computing Service| Public cloud (AWS) | Elastic Slurm clusters, managed updates and scaling |
| Lambda Managed Slurm | 1‑Click GPU clusters | AI‑tuned SLURM, monitoring, support via SchedMD |

Let me know if you'd like pricing, feature comparisons, or help selecting the best fit for your specific workload.