How Metaflow helps operate data science projects at scale


Details
A conversation with Ville Tuulos, former ML infrastructure manager at Netflix, and one of the main people behind Metaflow.
Hosted by Itai Bar-Sinai, Co-Founder and CPO at monalabs.io.
Synopsis:
Metaflow was started at Netflix in 2017 to address the needs of hundreds of ML projects ranging from NLP and computer vision to classical statistics. Today, over 70% of all data science projects at Netflix are powered by Metaflow. The project's rapid growth was fueled by our focus on user experience: Metaflow makes it easy for ML practitioners to quickly take their projects from prototype to production by leveraging existing cloud infrastructure. In this talk, Ville shares details about why Netflix built Metaflow, how teams use in practice, and how it is positioned in the rapidly evolving landscape of ML infrastructure tools.

How Metaflow helps operate data science projects at scale