Library

Managing Real-Life Data Science Projects with Metaflow

Sam recently had the pleasure of sitting down for a [conversation](https://twimlai.com/twiml-talk-326-metaflow-a-human-centric-framework-for-data-science-with-ville-tuulos/) with [Ville Tuulos](https://twimlai.com/network/ville-tuulos/), Manager of Machine Learning Infrastructure at Netflix. They spoke just hours after he announced the open-sourcing of their Python library, [Metaflow](https://metaflow.org/). At the time, we’d planned to launch a new video series called TWIML Democasts as a way to walk through concrete technologies in a visual way that the podcast format doesn’t allow. During the interview, Sam threw out the suggestion that if there was enough interest, we’d record a Democast for Metaflow. Well, the people have spoken and that time has come! We’re excited to share our third ever Democast, in which Ville gives us an in-depth, end-to-end demo of the entire Metaflow platform, including: - The Philosophy of Metaflow - Metaflow’s Tutorial School - Metaflow’s Relationship to Netflix’s Papermill - Workflow Execution Systems - Scalability And much more! We encourage you to check out Sam’s original [interview](https://twimlai.com/twiml-talk-326-metaflow-a-human-centric-framework-for-data-science-with-ville-tuulos/) with Ville, and of course, the Metaflow [documentation](https://docs.metaflow.org/introduction/what-is-metaflow) is a great resource to follow along with while you’re watching the video.