The Kale component of MiniKF and Arrikto Enterprise Kubeflow (EKF), allows you to operationalize your machine learning workflows as pipeline runs on Kubernetes deployments. Kale is provided as both a Python SDK and a JupyterLab extension.
In this section, we will describe how to use the Kale user interface (UI) from a Jupyter Notebook. The Kale UI enables you to annotate cells in a Notebook to define the steps of a machine learning pipeline. Using these annotations, Kale adapts your Python code so that the steps of your pipeline can be containerized and run on Kubernetes clusters.