RETRIEVAL-AUGMENTED GENERATION (RAG)

PRACTICE LAB CONNECTION In the Resume Review Agent lab, you will add a RAG step to the job-posting search tool, which you will first build in Chapter 65 and extend in Chapter 66. Instead of passing every job posting to the model, the harness will embed the postings, store them, and retrieve only the postings most relevant to a given resume or query. Decide how to chunk the postings, for instance by section such as responsibilities and requirements, and what metadata to attach, such as job title and date. Build a small test set of queries with the postings that should match, and measure precision and recall. Then use the prompts below.