A framework for real-life data science
Metaflow makes it quick and easy to build and manage real-life data science projects.
Metaflow is built for data scientists, not just for machines
Successful data science projects are delivered by data scientists who can build, improve, and operate end-to-end workflows independently, focusing more on data science, less on engineering.
Model with your favorite tools
Use Metaflow with your favorite data science libraries, such as Tensorflow or SciKit Learn, and write your models in idiomatic Python code with not much new to learn. Metaflow also supports the R language.
Build with Metaflow
Metaflow helps you design your workflow, run it at scale, and deploy it to production. It versions and tracks all your experiments and data automatically. It allows you to inspect results easily in notebooks.
Use with AWS or Kubernetes
When you need more scale, Metaflow provides built-in integrations to storage, compute, and machine learning services for AWS and Kubernetes. No code changes required.
Battle-hardened at Netflix
Metaflow was originally developed at Netflix to address the needs of its data scientists who work on demanding real-life data science projects. Netflix open-sourced Metaflow in 2019.
Release Highlights
Jul 27th Metaflow API docs published A complete reference guide to all Metaflow APIs.
Jun 21st Metaflow tags are now mutable Set up effective human workflows around Metaflow projects.
Jun 1st New Metaflow Resources A library of practial tips & how-to guides for Metaflow users.
Apr 25th Metaflow for Kubernetes and Argo Workflows A full data science stack for Kubernetes.
Jan 25th Metaflow Cards released Visualize results of workflows with a few lines of Python.
Get Started!
To get started quickly, follow the tutorial.
For an overview of Metaflow, see documentation for Python or documentation for R.
To learn how to deploy and operate Metaflow at scale, see Administrator's Guide to Metaflow.
Questions? Reach us at our chat room or by email at help@metaflow.org.