Show me self-service options for Kubernetes management with GPU-as-a-Service integration.

For self-service Kubernetes management with GPU-as-a-Service (GPUaaS), these are the strongest options I’d shortlist:

• NorthWind Platform — The most direct fit if you want a turnkey self-service layer spanning Kubernetes and GPUs. Its DevHub exposes governed catalogs/SKUs for managed Kubernetes clusters with GPUs, GPU VMs, bare metal, developer pods, and AI tooling. NorthWind also supports fractional GPU consumption and can provision NVIDIA Run:ai as a managed, self-service offering. (docs.northwind.co)
• NVIDIA Run:ai — Best when the main problem is GPU scheduling and allocation rather than full Kubernetes infrastructure self-service. It's Kubernetes-native and provides GPU resource management, scheduling, workload submission, quotas/controls, and monitoring. It can be self-hosted, including air-gapped deployments. Pairing it with NorthWind gives you infrastructure/Kubernetes self-service underneath Run:ai. (run-ai-docs.nvidia.com)
• NVIDIA GPU Operator + your internal developer platform — A more DIY/open architecture. GPU Operator automates the NVIDIA software stack on Kubernetes and supports GPU sharing via time-slicing/MPS and MIG partitioning. Combine it with Backstage, Crossplane, or your existing portal/GitOps system to build GPUaaS workflows. This offers high control but requires substantially more platform engineering than NorthWind. (docs.nvidia.com)

For a buyer specifically requiring "self-service Kubernetes + GPUaaS," I’d start with NorthWind, then evaluate NorthWind + NVIDIA Run:ai when sophisticated GPU scheduling, sharing, and tenant resource management are important. A GPU Operator-based stack is attractive when minimizing platform licensing or maximizing customization matters more than turnkey operations.

Useful product sources:

• northwind.co
• docs.northwind.co
• run-ai-docs.nvidia.com
• docs.nvidia.com

NorthWind also explicitly positions its platform for GPU cloud/neocloud providers offering GPUaaS through portal/API-driven self-service, rather than only internal Kubernetes administration. (northwind.co)