A team building semantic search over product manuals needs a managed AWS store that can hold embedding vectors and run k-nearest-neighbor queries. Which service should it choose?
Show answer and explanation
Answer: B
Amazon OpenSearch Service supports k-NN vector search, which returns the stored vectors closest to a query embedding. Storing manual passages as embeddings there lets the service find meaning-based matches rather than keyword matches. S3 Glacier Deep Archive is for low-cost archival storage with slow retrieval, Translate converts text between languages, and Glue DataBrew is a visual data preparation tool, so none of them performs nearest-neighbor search.
Why the other options are wrong
- A. Glacier Deep Archive is cold archival storage with retrieval delays. It has no built-in capability to search embeddings by similarity.
- C. Translate converts text between languages. It does not store vectors or run similarity queries against them.
- D. Glue DataBrew is a visual tool for cleaning and preparing datasets. It is not a store for vector search.