Hopsworks Feature Store after 4 years: Lessons learned and what's next

Tuesday, October 11, 12:15pm - 12:40pm (EDT)

Time shown in-04:00 America, New York

Speakers: Fabio Buso and Moritz Meister, Hopsworks

It’s 2022 and everyone has heard of Feature Stores. We started building Hopsworks Feature Store in 2018, as such, we have had many conversations with organizations who are deciding between building their own feature store and buying one. Given the increasing interest in building feature stores, we want to share the lessons learnt over the past four years and what lies ahead.
The lessons learned span from API design, architectural decisions to accommodate real-time use cases but at the same time support Data Scientists’ favorite tool - Python.

We will also look at the value of a feature store in building operational services around models fed by feature stores. The costs of integration of feature stores with model serving and feature logging is still placed on the developer. To ease the integration burden, Hopsworks now supports KServe for model serving. By choosing best-of-breed technologies, KServe and the Feature Store, we will show you how to take your MLOps to the next level.

Add to Calendar 2022/10/11 09:15:00 2022/10/11 09:40:00 America/Los_Angeles Hopsworks Feature Store after 4 years: Lessons learned and what's next Speakers: Fabio Buso and Moritz Meister, Hopsworks

It’s 2022 and everyone has heard of Feature Stores. We started building Hopsworks Feature Store in 2018, as such, we have had many conversations with organizations who are deciding between building their own feature store and buying one. Given the increasing interest in building feature stores, we want to share the lessons learnt over the past four years and what lies ahead.
The lessons learned span from API design, architectural decisions to accommodate real-time use cases but at the same time support Data Scientists’ favorite tool - Python.

We will also look at the value of a feature store in building operational services around models fed by feature stores. The costs of integration of feature stores with model serving and feature logging is still placed on the developer. To ease the integration burden, Hopsworks now supports KServe for model serving. By choosing best-of-breed technologies, KServe and the Feature Store, we will show you how to take your MLOps to the next level.

Online false MM/DD/YYYY 30 OPAQUE apQfLtmRnzOiFbgoNmal132273

Online

Feature Store Summit 2022

Link: