Tuesday, October 12, 12:20pm - 12:50pm (EDT)
This presentation is part of the Feature Store Summit.
---
The value of a Feature Store lies in the reduced time to production for new ML models. This talk will focus on how to leverage the Hopsworks Feature Store to compute and ingest features and make them available to operational models making real-time predictions, with low latency and preventing inconsistency between the training and serving features.
To do this at scale we leverage four pillars: Accessibility through a range of environments like Spark, Python, SQL and Flink. A scale out metadata layer to discover, govern, quality assure and reuse features. A scalable and reliable ingestion engine to guarantee feature freshness and consistency. And RonDB, the world’s fastest key-value store backing not only the Online Feature Store but also handling all metadata generated within Hopsworks.
You're going to "Fast, Fresh Data for AI at Scale with the Hopsworks Feature Store".
We've sent a confirmation email to your email address. Be sure to check your junk folder in case you haven't received the confirmation.
You're interested in "Fast, Fresh Data for AI at Scale with the Hopsworks Feature Store".
We've sent a confirmation email to your email address. Be sure to check your junk folder in case you haven't received the confirmation.
Thank you!
Your changes have been saved. Thanks for keeping us updated.
https://us02web.zoom.us/j/84897688570
Featurestore.org, featurestoreorg@gmail.com