
ML platform teams, features stores, versioning in data pipelines, and where MLOps extends DevOps with Aurimas Griciūnas and Piotr Niedźwiedź
ML Platform Podcast
00:00
What Is a Feature Store?
A feature store is a library or framework on top of storage, but it has to have two sites ingestion and server. Feature store itself serves, not serves, but stores the features that it has already ingested. So when it comes to ingesting from outside sources, as mentioned, like databases or third party APIs, this requires a data engineering flow. And it's less machine learning problem, more a data engineering problem.
Play episode from 29:13
Transcript


