A payments company produces many builds of its fraud scoring model on Vertex AI. Auditors want a record of every build, the one currently serving live traffic, and the evaluation numbers for each build. Which capability should the team use?
Show answer and explanation
Answer: A
Vertex AI Model Registry stores model versions, lets teams attach aliases such as production to the version that serves traffic, and keeps evaluation metrics with each version so they can be compared. It gives auditors a central, consistent record. The other options are manual or runtime-focused and cannot reliably track which version is live along with its lineage.
Why the other options are wrong
- B. A spreadsheet relies on people updating it by hand. It has no link to the actual model artifacts or to which version is deployed.
- C. A hand-filled BigQuery table can work, but it duplicates registry functions manually and is prone to missing or stale entries.
- D. Monitoring tracks runtime signals such as latency and custom metrics. It does not manage model versions or record which artifact is in production.