Article Source

• When to use Kubernetes natively over Kubeflow for ML

When to use Kubernetes natively over Kubeflow for ML

The ZenML Meet The Community event happens every week on Wednesdays 9AM PT / 5 PM CEST (Register here: https://www.eventbrite.de/e/zenml-mee…. While the main goal of the meeting is to chat between each other about latest features and bugs, the ZenML core team sometimes does a demo of new features. Therefore, we sporadically record these meetings and share with a wider audience.

The last meeting was a chance to showcase the latest #Kubernetes native orchestrator for the first time. The Kubernetes orchestrator is to be used as a light-weight alternative to more heavy duty orchestrators like #Airflow or #Kubeflow.

About ZenML:

ZenML is an extensible, open-source MLOps framework for creating portable, production-ready MLOps pipelines. Built to enable collaboration among data scientists, ML Engineers, and MLOps Developers, it has a simple, flexible syntax, is cloud- and tool-agnostic, and has interfaces/abstractions that are thoughtfully designed for ML workflows.

🔗 Code Repository: https://github.com/zenml-io/zenml/tree/main/examples/kubernetes_orchestration

🔗 MLOps Platform Sandbox: https://sandbox.zenml.io

🔗 Slack community: https://zenml.io/slack

🔗 ZenML Website: https://zenml.io