Feast Community Newsletter #4

[Danny Chiao]

Nov 8, 2021

Announcements

Feast 0.15

We released a new version of Feast! A big thanks to those contributing to the release. This release introduces: [Experimental] Push-based streaming support (docs): users can push streaming data into online stores for fresher features at serving time. Entity aliasing (docs): improved support for cases when an entity (e.g. "user") has different column names in different source tables (e.g. "spammer", "reporter"). Feature Transformation Server: a server that executes on demand transformations. The existing feature server (e.g. deployed with AWS Lambda) executes on demand transformations already. This new server integrates with Feast Serving (java server) for latency sensitive use cases. Easy way to test offline/online store plugins using the existing Feast test suite. See docs for details.  If you have questions on any of these, please reach out to the Feast team!

Blog posts and videos

Check out some new content featuring Feast: [Blog] Bringing Feature Stores to Azure (post) [Blog] Key Takeaways from the First Feature Store Summit (post) [Blog] Bring ML Close to Data Using Feast and Flyte (post) [Podcast] Build Composable and Reusable Feature Engineering Pipelines (podcast)

Upcoming events

Times below are in Pacific time unless otherwise noted:  Nov 9 @ 10AM: Feast community call (agenda, zoom link) Please add agenda items to discuss! Current topics include: Profiling Feast serving and on demand feature views Nov 18 @ 5PM EST: NYC community happy hour (event, pw: tectonhh2021)

Community contribution highlights

Bug fixes for third party integrations / plugins Stream ingestion and Redis pipelining (@vas28r13) Improve performance of _convert_arrow_to_proto (@nossrannug) Bug fixes with Pandas types in on demand feature views (@judahrand) WIP Trino plugin (@MattDelac)

Work in progress

Low latency serving

The team has profiled a sample scenario in using Feast Serving with on demand feature views for large batch sizes and found action items to improve serving performance. See details in On-Demand Transformations Latency Investigation.   RFC-025: Online JVM Client RFC: Low Latency Serving

Miscellaneous

Other noteworthy ongoing work without RFCs: Documentation on best practices for deploying Feast (@pyalex) Data quality monitoring (early investigations, @adchia, @pyalex) Porting of Feast Java into main repo (@adchia) Scalable historical retrieval (@MattDelac, @mavysavydav) Snowflake offline store plugin (@milescadkins) Astra DB online store plugin (@hamzakpt) Alibaba Cloud plugin (@777ki) Trino offline store plugin (@MattDelac)

Join the conversation on Slack, give feedback on our roadmap, or checkout our GitHub!