MODULE 5 · DAY 2
Docs Assistant — Naive RAG
Grounding an LLM in your own documents
Gourav Shah · School of DevOps & AI · Hands-on
M5·01
What you'll learn
The anatomy of a real GenAI app — and how to wire it in containers.
M5·02
A model alone can't answer about YOUR docs
It answers confidently from training data that predates your runbooks.
M5·03
Anatomy of a GenAI application
Four parts, wired together. You already built the first two in M2 and M3.
M5·04
A vector DB shelves by meaning, not title
A filing cabinet needs the exact name. A librarian understands what you mean.
M5·05
Embeddings: text becomes coordinates
An embedding model places each passage as coordinates in meaning-space.
M5·06
The naive-RAG pipeline
Ingest once. Then answer every question with the same embedding step.
M5·07
Where the pieces run
Models stay native for Metal; the vector store and app are containers.
M5·08
ChromaDB — the lightest vector store
Zero-config and under 2 GB — the right default for a 16 GB laptop.
M5·09
Wired via environment variables
Hand-author the compose service by service — endpoints are just env vars.
M5·10
Learning Mode — watch the pipeline run
The app surfaces each step live — the invisible parts of RAG become visible.
M5·11
Where naive RAG breaks down
It works when the question matches the wording. It fails in predictable ways.
M5·12
TO THE LAB
Retrieve, then generate — grounded.
Hand-author the compose, ingest Acme's runbooks, watch Learning Mode explain.
Next: Lab — build the Docs Assistant. · Gourav Shah · School of DevOps & AI
M5·13