Question 190
At fictional Greyson Customer Network, a Snowflake data and AI engineering team is conducting an accountable generative AI operations change with regional service policies and accountable stakeholders. The separately composed technical incident is: A data engineer changes embedding settings for a retrieval service, and the new vectors cannot be inserted because the stored VECTOR column expects a different dimension length. Assume only these documented facts and use safe, authorized production-change procedures; exact current Snowflake feature syntax and region or account availability require current review. Which repeatable assurance or control is most relevant to preventing this situation?
Explanation
Writer-proposed answer: B (UNVERIFIED)
Preliminary individual option explanations — NOT expert validated
Correct Answer
Validate vector dimensions in CI, version embeddings and require a controlled schema-migration plan for all model upgrades. Vector data types and operations require dimensional compatibility even before semantic quality and model consistency are evaluated.
Incorrect Answers
This response concerns Ambiguous product review classification, which is different from the stated failure. A data engineer changes embedding settings for a retrieval service, and the new vectors cannot be inserted because the stored VECTOR column expects a different dimension length.
Custom inference model deployment parity is not the cause of the reported failure: A data engineer changes embedding settings for a retrieval service, and the new vectors cannot be inserted because the stored VECTOR column expects a different dimension length.
Unauthorized tenant content in retrieval alone does not establish the needed evidence: New vectors insert with the expected shape and retrieval tests find the intended records without dimension conversion or silent truncation.