RAG Lab · zero-ml

RAG LabModule 01 · see it, type it, ship it

Curious0 XP

0 / 51 stamps0 of 17 learned

Seventeen techniques from the freeCodeCamp course by Lance Martin. Each one is a short loop: watch it work, predict what happens, build it by typing the lines, read the full code from the course notebook, then take the final touch, the one call you reuse.

The loop0/4

Index, retrieve, generate. The three moves every RAG app makes.

01Overview0:00 02Indexing5:53 03Retrieval10:40 04Generation15:52

Rewrite the question0/5

Users ask badly. Change the question until the right chunk shows up.

05Multi-Query22:14 06RAG Fusion28:20 07Decomposition33:57 08Step Back40:31 09HyDE47:24

Send it right0/2

Pick the right store, and turn plain words into a real query.

10Routing52:07 11Query Construction59:08

Index smarter0/3

Change what you store, so retrieval has better material.

12Multi-Representation1:05:05 13RAPTOR1:11:39 14ColBERT1:19:19

Check yourself0/3

Grade the retrieval, correct course, and ask where RAG goes next.

15CRAG1:26:32 16Adaptive RAG1:44:09 17Future of RAG2:12:02

Coming nextMachine learning performance measures

The next module, with the same loop: predict, build it, see it move.

Module 01 · RAG From Scratch

RAG From Scratch

Learn AI by watching it work. Start with RAG from scratch.

Pick a section on the map, or start at the beginning. Your progress is saved on this device.

Start with the overview →

Based on the course by Lance Martin (LangChain) · watch on YouTube