Developing Feast, the Leading Open Source Feature Store, with Willem Pienaar (Gojek, Tecton)

09/03/2021 1h 11min Episodio 24
Developing Feast, the Leading Open Source Feature Store, with Willem Pienaar (Gojek, Tecton)

Listen "Developing Feast, the Leading Open Source Feature Store, with Willem Pienaar (Gojek, Tecton)"

Episode Synopsis

Willem Pienaar is the co-creator of Feast, the leading open source feature store, which he leads the development of as a tech lead at Tecton. Previously, he led the ML platform team at Gojek, a super-app in Southeast Asia.Learn more:https://twitter.com/willpienaarhttps://feast.dev/Every Thursday I send out the most useful things I’ve learned, curated specifically for the busy machine learning engineer. Sign up here: https://www.cyou.ai/newsletterFollow Charlie on Twitter: https://twitter.com/CharlieYouAISubscribe to ML Engineered: https://mlengineered.com/listenComments? Questions? Submit them here: http://bit.ly/mle-surveyTake the Giving What We Can Pledge: https://www.givingwhatwecan.org/Timestamps:02:15 How Willem got started in computer science03:40 Paying for college by starting an ISP05:25 Willem's experience creating Gojek's ML platform21:45 Issues faced that led to the creation of Feast26:45 Lessons learned building Feast33:45 Integrating Feast with data quality monitoring tools40:10 What it looks like for a team to adopt Feast44:20 Feast's current integrations and future roadmap46:05 How a data scientist would use Feast when creating a model49:40 How the feature store pattern handles DAGs of models52:00 Priorities for a startup's data infrastructure55:00 Integrating with Amundsen, Lyft's data catalog57:15 The evolution of data and MLOps tool standards for interoperability01:01:35 Other tools in the modern data stack01:04:30 The interplay between open and closed source offeringsLinks:Feast's GithubGojek Data Science BlogData Build Tool (DBT)Tensorflow Data Validation (TFDV)A State of FeastGoogle BigQueryLyft AmundsenCortexKubeflowMLFlow

More episodes of the podcast Machine Learning Engineered