Articles for people who
ship the stack
Rewrites on React, Node.js, TypeScript and AI. A translated body is shown when it exists. If it does not, you see the English text and a note that says so.
Tagged: BM25
Fifty Years of Retrieval Research Hiding Inside Your RAG Pipeline
RAG chunking, hybrid search, HyDE, reranking and evals all have decades-old ancestors in information retrieval. A tour of the research and where it falls short.
17 min readDebugging RAG Retrieval: Recall@K, BM25, Hybrid Fusion and Reranking
A small experiment on a RAG pipeline that ranked the right chunk fourth, showing how chunking, BM25, rank fusion and reranking change Recall@3 and MRR.
11 min readEvaluating RAG Retrieval Beyond ANN Recall, with Turbopuffer as the Example
A fast vector index can still miss the answer. Learn Turbopuffer's storage tradeoffs, hybrid BM25 plus vector search, and how a golden set measures real recall.
7 min readHybrid Retrieval for RAG: BM25, Rank Fusion and Reranking Beat Vectors
Pure vector search misses exact terms. BM25 plus embeddings, Reciprocal Rank Fusion and a cross-encoder reranker took this RAG setup to 91% retrieval accuracy.
6 min readContextual Retrieval for RAG: What the 67% Figure Hides, What It Costs and When It Fails
Contextual retrieval for RAG: how situating prefixes fix orphaned chunks, why the 67% figure stacks three changes, what prompt caching saves, and when it fails.
13 min readFive Ways to Index Vectors for RAG, Tested on a Million Wikipedia Embeddings
Five ways to index vectors for RAG tested on a million Wikipedia embeddings: flat, IVFFlat, HNSW, BM25 hybrid and DiskANN, with latency and recall numbers.
9 min readRetrieval Over an Engineering Wiki: BM25, Token-Safe Dense Windows and RRF
Evaluation-first retrieval for a technical wiki: exact-evidence targets, a BM25 baseline, a MiniLM truncation bug, token windows and reciprocal rank fusion results.
7 min readHybrid Search for RAG: BM25, Vectors and Reciprocal Rank Fusion in Practice
Hybrid search for RAG: why vector-only retrieval loses exact IDs, how Reciprocal Rank Fusion merges BM25 with embeddings, and where hybrid still falls short.
11 min readThe RAG Retrieval Toolkit: BM25, Dense, HyDE, RRF, MMR and Reranking
The RAG retrieval toolkit: BM25, dense vectors, HyDE, RRF, MMR and reranking, when to use each, rough latency per stage, and how a hybrid pipeline fits them.
8 min read
About these articles
Request a 24h estimate
Need the same stack in a production operator layer? Send the brief — estimate within 24 hours.