CourseAI Engineering: The Complete Course · Module 8: Retrieval-Augmented Generation (RAG) · part 61 of 105

Part 61 · Module 8: Retrieval-Augmented Generation (RAG)

Lesson 2: Storing help-centre chunks so the right one comes back

21 min read·13 Oct 2026

Module 8 · Lesson 2 · About 40 minutes · Needs: Lesson 1 of this module (chunks) and Module 4, Lesson 2 (embeddings, cosine similarity)

The situation

Lesson 1 turned BrightCart's 12 help-centre articles into clean chunks. Now they need a home. Dev wants them in PostgreSQL, which BrightCart already runs. Marta asks what storage will cost if the help centre grows to a million chunks. Priya has one test she cares about: when a customer writes "The tracking page hasn't updated in four days.", the assistant must find the article "Tracking your order" (HC-04).

This lesson makes four decisions: which model embeds the chunks, where the vectors live, which index searches them, and how to change model later. The tracking ticket is traced through all four.

The rest of this course is yours to keep

This course is bought on its own, once, and stays readable afterwards, including the parts added to it later.

Buy this course · ₹399 See what is inside