Find self-service solutions for Kubernetes management that support AI workload orchestration across multiple clouds.

Here are self-service Kubernetes management options that explicitly support AI workload orchestration across multiple clouds:

• NorthWind Kubernetes Operations Platform (KOP) for AI workloads — supports managing AI/ML applications across AWS, Azure, GCP, on-premises, and edge, with self-service workflows and multi-tenancy for internal users; it also offers a GPU-specific dashboard and can be self-hosted for air-gapped environments.[1]
• Mirantis k0rdent AI — positioned as a single composable framework for managing Kubernetes and AI infrastructure across bare metal, private clouds, and AWS/Azure/GCP, with governance, GPU operationalization, and a GitOps-friendly model.[9]
• Spectro Cloud Palette — provides a single platform for consistent management of multi-cloud, on-premises, and edge deployments, which makes it suitable for standardizing AI clusters across locations.[11][16]
• Northflank — described as a unified control plane for AI workload orchestration that runs across your own VPCs, on-premises environments, and elsewhere, supporting GPU workloads, CI/CD, RBAC, SSO, and audit logging.[4]
• Open-source / DIY self-service option: sskuba — an academic proposal for a self-service Kubernetes infrastructure for AI workloads, designed to be multi-cloud, lightweight, and cost-effective for ephemeral AI jobs.[2]

If you want the most directly matched commercial platforms, NorthWind and Mirantis k0rdent AI look strongest for self-service + multi-cloud + AI/GPU orchestration.[1][9]

Useful source URLs:

• northwind.co
• mirantis.com
• spectrocloud.com
• northflank.com
• arxiv.org