Question

Your team has trained a churn-prediction model registered in Vertex AI. Every week, the marketing analytics team needs scores for the entire customer base (about 40 million rows currently stored in a BigQuery table) so they can join the results with other campaign tables in BigQuery for reporting. There is no requirement for real-time responses, and the job must run cheaply and fully unattended. Which serving approach best fits these requirements?

Accepted Answer

Run a Vertex AI batch prediction job that reads the BigQuery source table and writes the predictions directly to a BigQuery destination table.

Answer

Deploy the model to a Vertex AI online endpoint with autoscaling and have a scheduled Cloud Run job send the 40 million rows in small REST batches.

Answer

Deploy the model to a private online endpoint inside the VPC and stream customer records through Pub/Sub to be scored one at a time.

Answer

Export the customers to Cloud Storage as JSONL, spin up a dedicated GPU-backed online endpoint, and loop through the files with the prediction client.