Kale 2.0 has been officially released as part of the Kubeflow ecosystem. Originally a graduate thesis project turned startup product, Kale enables data scientists to build and run ML pipelines directly from Jupyter notebooks. The 2.0 release brings full KFP v2 compatibility with a rewritten backend, a modernized JupyterLab 4 extension, a new documentation site, and usability improvements. The author reflects on the project's revival and outlines its future direction: deeper integration with the Kubeflow SDK and JupyterLab's evolving ecosystem, including AI-assisted pipeline authoring via Jupyter AI. A Google Summer of Code 2026 project will explore low-code, in-notebook orchestration of multiple notebooks.
Table of contents
A graduate project, finally homeWhat’s in 2.0Why Kale, still?Thank youWant to Contribute?Sort: