Question

Your data science team maintains a BigQuery table that is refreshed daily by a Dataflow pipeline with customer-level aggregated features. Several downstream teams want to retrieve the latest feature values with low-latency online lookups for their real-time recommendation services. You want to expose these features from the existing BigQuery source with minimal engineering effort and keep the online store automatically up to date after each daily refresh. Which approach should you use?

Accepted Answer

Create a Vertex AI Feature Store (with the new BigQuery-backed FeatureView) that references the BigQuery table and configure a scheduled sync to update the online store daily.

Answer

Export the BigQuery table to Cloud Storage each day and have each downstream team load the CSV into an in-memory cache within their serving application.

Answer

Write a custom Cloud Run service that queries the BigQuery table directly on every online prediction request to fetch feature values.

Answer

Set up a Dataflow streaming job that continuously writes each row into Cloud Bigtable and have teams query Bigtable through their own client libraries.