Sample question 1 — Applications of Foundation Models
Incident tickets at Riverbend Cable Ops keep recurring around the same theme: an analytics team wants embeddings for product catalog search using a managed AWS model API. The security engineer for AI workloads should institute the lasting AWS AI Practitioner (AIF-C01) fix rather than another one-off patch. How should they correct the situation?
- A. Call a Bedrock embedding model to vectorize catalog text for similarity search.
- B. Skip vectors and rely only on SQL LIKE predicates for semantic search.
- C. Train embeddings by averaging ASCII codes of product names manually.
- D. Use a text generation model temperature setting as if it were an embedding vector.
Answer: A. Option A is correct: "Call a Bedrock embedding model to vectorize catalog text for similarity search." Bedrock embedding models produce vectors suited for similarity search. Option D is incorrect because Temperature is not an embedding vector. Option B is incorrect because SQL LIKE lacks semantic similarity. Option C is incorrect because ASCII averages are not learned embeddings.