Day 1
Oct 12, 2021
10:30 am

Feature Store: The Heart of Your Operational ML Pipeline

Yaron Haviv from Iguazio talks how the feature store supports real-time and batch use cases across training and serving environments.

About this session

Feature stores accelerate the development and deployment of AI applications by automating feature engineering.  They provide a single pane of glass to build, share and manage ML features across projects and teams. Advanced feature stores tap into production data and online or real-time event sources.  They then run a set of analytical or statistical transformations on the ingested data. This way, they create an offline dataset for training, a real-time dataset for serving, and statistical data analysis for monitoring model accuracy and drift.

The feature store is the center piece in every ML infrastructure.  When used properly, it can solve the most challenging part of operationalizing ML pipelines:  Producing the right data for all ML applications and stages.  Maximizing the feature store’s value requires tight and glue-less integration with model training, serving and monitoring frameworks.

Join this session to hear how feature stores can function as the heart of your operational ML pipeline, supporting real-time and batch use cases across training and serving environments. Discover how they accelerate your path to production and eliminate silos between data science, data engineering and ML engineering teams.

Add to Calendar

Moderator

Session Speaker

Session Speaker

Session Speaker

Session Speaker

Session Speaker

Join our Slack channel to stay up to date on all the latest feature store news, including early notification when the conference details emerge.