60 Days of Production AI Systems · chapter 3 of 10
Retrieval Foundations
Chunking, embeddings, vector stores, keyword and hybrid search, and the permissions retrieval must respect.
6 of 6 published
Chapter guide · 7 min readNo model can reason about evidence your retriever never returnedThe whole chapter in one read. Then go part by part below.- 01
Day 13 · Explainer · 2 min read
Why chunk boundaries shape answer quality
How chunk boundaries decide what evidence the system can actually retrieve.
On the map: Ingestion - 02
Day 14 · Explainer · 2 min read
Why embeddings are useful and easy to overtrust
Why semantic similarity is useful, but not the same as correctness or trust.
On the map: Ingestion, Vector Index - 03
Day 15 · Explainer · 2 min read
Why vector stores are infrastructure, not magic memory
Why vector databases need to be operated like infrastructure, not treated like magic memory.
On the map: Vector Index - 04
Day 16 · Explainer · 2 min read
Why keyword search and semantic search both matter
Why exact words and semantic meaning both matter in production retrieval.
On the map: Retrieval - 05
Day 17 · Explainer · 2 min read
Why hybrid search is often the practical default
Why combining retrieval signals is often more practical than betting on one search method.
On the map: Retrieval - 06
Day 18 · Explainer · 2 min read
Why relevant data can still be unauthorized data
Why relevant evidence is still wrong evidence if the user was not allowed to see it.