Canonical, the publisher of Ubuntu, has announced the general availability of Charmed Kubeflow 1.8. Representing a significant advancement in AI and ML development and deployment, Charmed Kubeflow 1.8 is an open-source, end-to-end MLOps platform. It allows professionals to develop and deploy AI/ML models conveniently and supports multiple environments, including cloud, hybrid, or multi-cloud scenarios. Crucially, it also facilitates running AI/ML workloads in air-gapped environments.
A challenge often encountered in MLOps platforms is the need for network connections, which can pose security and compliance issues for specific organisations. Yet, Charmed Kubeflow overcomes this challenge by facilitating offline workload execution in air-gapped environments, in addition to public clouds and on-premises data centres.
With this new feature, the platform provides enhanced security for projects dealing with sensitive data, particularly those within highly regulated industries. As such, most machine learning workflow can be completed within a single tool, eliminating the need for time-consuming tool connection and compatibility assurance.
Charmed Kubeflow 1.8 also recognises the diverse needs of different AI projects, presenting enhanced capabilities for end users to customise their MLOps tools. This allows integration of any image within their Jupyter Notebook, enabling professionals to choose desired tools and libraries while focusing on developing machine learning models rather than tooling maintenance.
Moreover, tools or components can be plugged in or out based on the use case, ensuring effective work practices. This differentiating feature allows organisations to move beyond experimentation with Canonical's supported solution, with the freedom to add their own Notebook images and develop models.
Kubeflow, originally designed to run AI at scale, encapsulates the entire machine learning lifecycle within a single tool. At the heart of this project lie the Kubeflow Pipelines, which specialise in automating machine learning workloads. These pipelines are a core reason organisations seeking to scale AI projects prefer Kubeflow.
Charmed Kubeflow now leverages the newly introduced Kubeflow Pipelines 2.0, further simplifying the automation process and facilitating smoother migrations, providing enterprise support, security patching, and timely bug fixes.
Kimonas Sotirchos, Working Group Lead in the Kubeflow Community, says: "I'm thrilled to be part of the upstream community's Kubeflow 1.8 release and proud of the Charmed Kubeflow team for driving the release as well as providing feedback along the way."
"Charmed Kubeflow 1.8 is a great way for newcomers and experienced users to try out all the latest and most significant features in Kubeflow, like KFP V2 and PVC browsing."
Charmed Kubeflow 1.8 serves as the foundation of a dynamic ecosystem catering to AI projects, with the integration of leading open-source tools such as Charmed MLflow, KServe, and Seldon. This ecosystem ensures a comprehensive solution for AI projects, from experiment tracking to model serving. Specifically, the integration with Charmed MLflow facilitates experiment tracking and model registry.
This lightweight machine learning platform allows professionals to quickly start projects locally or on the public cloud and then easily migrate to a fully integrated, open-source solution.