Question

Three ML teams at your company each maintain their own preprocessing pipelines that compute nearly identical customer-level features (e.g., 30-day purchase frequency, average session length). This leads to duplicated engineering effort, inconsistent definitions, and drifting values across teams. Leadership wants a governed, reusable approach so features are defined once and shared for both training and low-latency online serving. Which approach best addresses this?

Accepted Answer

Centralize the shared features in Vertex AI Feature Store, organizing them into feature groups/entity types that all teams read from for training and online serving.

Answer

Have each team publish its computed features to a shared BigQuery dataset and query it directly at serving time for every prediction request.

Answer

Store the features as CSV exports in a shared Cloud Storage bucket and have each team load the latest file before training.

Answer

Ask each team to copy the best-performing team's preprocessing notebook into their own project and run it independently.