Feast

Integrating Arize with open sourced lightweight feature store, Feast.

Arize and Feast are two platforms aimed at different, but connected, parts of the ML pipeline. Arize helps you visualize your model performance, understand drift & data quality issues, and share insights as your Evaluation Store. Feast (i.e, Feature Store) is an operational data system for managing and serving machine learning features to models in production.

Integration with Feast is simple in four steps.

Step 1: Log production by calling arize.log after materializing and store.get_online_features.

Step 2: At training time, log validation under batch_id of offlineafter fetch historical feature store.get_offline_features.

Step 3: Set-up match environment on Arize.

Step 4: Troubleshoot and observe any data inconsistency issue with Arize.

Tutorial Notebook

We have set up a very basic example using Feast with Arize.

Google Colaboratory