mediumAI Engineering

What is hybrid search and why combine BM25 with vector search?

Q: What is hybrid search and why combine BM25 with vector search?

**TL;DR.** Run **BM25 keyword search** and **vector search** in parallel and fuse the rankings (usually **Reciprocal Rank Fusion**). Vectors understand paraphrase but miss exact identifiers; BM25 nails exact terms but not meaning. Fusion raises recall on both query types — the default for serious RAG. query BM25 <text class='d-sub' x='206' y='56'

226 views

Understand the problem

Lexical + semantic retrieval fused with RRF: covering each method's blind spots.

hybrid-searchbm25rrfretrieval

Attempt it yourself

Sketch your approach before reading the solution — that's what interviews test.

Stuck? AI Nudge Available

Get a conceptual hint to guide your logic without spoiling the final implementation.

Study the solution

The solution is waiting

Give it an honest attempt first — then compare your thinking with the full walkthrough.

Read the code

Reciprocal Rank Fusion

function rrf(lists: string[][], k = 60): string[] {
  const score = new Map<string, number>();
  for (const list of lists)
    list.forEach((id, rank) =>
      score.set(id, (score.get(id) ?? 0) + 1 / (k + rank + 1)));
  return [...score.entries()].sort((a, b) => b[1] - a[1]).map(([id]) => id);
}

const fused = rrf([await bm25(q, 50), await vector(q, 50)]).slice(0, 20);

Join the discussion

Discussion (0)

No responses yet. Be the first to share what you think.

How do you handle PII and sensitive data when calling LLM APIs?

What are guardrails and how do you implement them around a model?