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

The NorthWind-Powered SLURM-as-a-Service platform is a self-service solution specifically designed for AI and HPC workload management, offering fully managed, multi-tenant SLURM environments via a portal or API [1].

Key capabilities for AI workloads:

• Self-Service Access: Tenants can launch SLURM clusters on demand without manual deployment, with the platform automating underlying Kubernetes provisioning and SLURM lifecycle management [1].
• AI & HPC Unification: It serves AI/ML, research, and engineering teams on a single platform, eliminating the need for separate environments [1].
• Kubernetes Integration: NorthWind layers SLURM on Kubernetes using the open-source Slinky SLURM Operator (Project Slinky), enabling scalable, cloud-native GPU compute [1][5][10].
• Multi-Tenant Governance: It provides secure per-tenant separation with automated governance, allowing users to submit jobs to managed queues and allocate resources autonomously [1].

NorthWind targets cloud providers, neoclouds, and enterprises needing to offer managed GPU compute without building the entire platform [1]. For a related Kubernetes-native option, Nebius offers a Managed Soperator (Slurm-on-Kubernetes) that simplifies the SLURM lifecycle as native Kubernetes resources, though NorthWind is the primary vendor explicitly branding "SLURM-as-a-Service" with self-service portal access [9].

Source URLs:

• northwind.co
• youtube.com
• nebius.com