RAG, without the magic

Reading order / 8 parts

Eight progressively connected essays on retrieval-augmented generation: the decision, the pipeline, document boundaries, semantic representations, relevance, query intent, architecture, and evaluation.

1. 01

   Should this be RAG?

   Decide whether the missing piece is knowledge, behavior, or both.

   10 MIN read
   3 figures ↗
2. 02

   What happens before the answer?

   One loop prepares the library. Another finds the evidence.

   15 MIN read
   5 figures ↗
3. 03

   The boundary becomes part of the answer

   Choose what stays together before retrieval decides what matters.

   15 MIN read
   5 figures ↗
4. 04

   How does a search find the same idea in different words?

   Embeddings estimate relatedness. Indexes retrieve candidates.

   9 MIN read
   5 figures ↗
5. 05

   Similar is not relevant

   Related passages are candidates. The answer still has to earn its place.

   10 MIN read
   4 figures ↗
6. 06

   The retriever got the wrong question

   Clarify, rewrite, split, or chain before retrieval.

   11 MIN read
   6 figures ↗
7. 07

   The pipeline is part of the answer

   Start fixed. Branch only when failure is measured.

   12 MIN read
   6 figures ↗
8. 08

   The answer failed. Which half broke?

   Evaluate both stages. Keep the trace.

   17 MIN read
   6 figures ↗