A Generative AI Engineer is building a RAG application for answering employee questions on company policies. What are the steps needed to build this RAG application and deploy it?

Question

Options

• AIngest documents from a source -> Index the documents and saves to Vector Search -> User
• BUser submits queries against an LLM -> Ingest documents from a source -> Index the documents
• CIngest documents from a source -> Index the documents and save to Vector Search -> Evaluate
• DIngest documents from a source -> Index the documents and save to Vector Search -> User

• A
  87% (13)
• C
  7% (1)
• D
  7% (1)

Explanation

The canonical RAG deployment pipeline is: (1) Ingest source documents, (2) chunk and index them into a Vector Search index, (3) accept a user query at runtime, (4) embed the query and retrieve relevant chunks, (5) pass chunks as context to the LLM to generate an answer, (6) return the answer to the user, and (7) deploy the assembled application. Option B starts with the user query before documents are ingested, which is logically impossible. Option C inserts an evaluation step in place of the live query-response loop, conflating offline evaluation with the runtime serving path. Option D is similar to A but omits critical evaluation steps before deployment, making A the most complete and correctly ordered description.

No community discussion yet for this question.