Kubeflow AI and MLOps at any scale
Enterprise-ready Charmed Kubeflow, the fully supported MLOps platform for any cloud. Charmed Kubeflow is Canonical's enterprise-ready MLOps platform. Deploy, scale, and manage AI workflows across clouds, VMs, or bare metal. A complete solution for sophisticated data science labs. Upgrades and security updates – all supported in the free, open source distribution.
Apply to join our preview for Managed Kubeflow on Azure! Learn more and sign up.
Why choose Charmed Kubeflow?
The fastest way to deploy production-ready Kubeflow, with full support and zero lock-in. Run anywhere, scale effortlessly, and empower your data scientists.
One solution, any cloud
Deploy Kubeflow on public clouds, private infrastructure, or air-gapped environments.
Multi-user collaboration
Enable secure workspaces with role-based access for data science teams.
Scale with confidence
Run parallel experiments at any scale, with GPU acceleration built-in.
Open and cost-efficient
No licensing fees. No usage limits. Backed by Canonical’s enterprise-grade support.
Predictable pricing for enterprise-grade Kubeflow
10 years security maintenance
Open source
Simple per node, per year, subscription
Scale experiments, effortlessly
Charmed Kubeflow lets you scale from a single-node lab setup to thousands of distributed training jobs. Built on Kubernetes, it offers native horizontal scalability and multi-cloud elasticity without extra complexity.
- Run distributed training jobs in parallel
- Leverage Kubernetes-native auto-scaling
- Separate training and inference environments easily
Optimized for every cloud
Charmed Kubeflow works seamlessly with major cloud Kubernetes services – including AWS, Azure, and GCP. Deploy with Juju for consistent, declarative operations across environments.
- GPU-ready configurations out of the box
- Support for AKS, EKS, GKE and more
- Built-in hybrid and multi-cloud support
Your full-stack MLOps platform
Accelerate time to value with a complete, modular MLOps stack. Combine Charmed Kubeflow with Charmed Kubernetes, Ceph, observability tools and more – all maintained by Canonical.
- Production-grade MLOps with full support
- Integrates with data lakes and inference engines
- Validated with leading OEMs and silicon vendors
Learn more about Charmed Kubeflow
Key considerations, benefits, the differences from the upstream project and how to get started with one of them.
This blog explains the environments Charmed Kubeflow can run in and how to deploy it. Learn how to approach deployment based on your specific use case, existing infrastructure, long-term strategy, and level of expertise.
This blog explores Kubeflow pipelines, use cases, components, benefits, and architecture.