Reciprocal rank fusion explained: exact words, vectors, and re-rankers merged into one ranked list
lilyraynyc · x · 2026-11-07
At Tech SEO Connect in Durham, Lily Ray presented slides explaining reciprocal rank fusion: a single search returns one fused ranked list combining exact-word matching, vector-based semantic meaning, and a re-ranking pass over the shortlist. Additional slides show one query routed by intent to a purpose-built index (e.g. troubleshooting, with events and trials sections) before rank fusion — a useful mental model of the modern AI search retrieval stack.
More from Apps
- ChatGPT consumer lead Adam Fry discusses what sets OpenAI's Dots agents apart — AarushSelvan · 2026-11-07
- Solo dev used Claude to build an in-game Studio mode that ships to players — Grobot93 · 2026-11-07
- Halai Play lets anyone build and ship games via MCP with Claude or ChatGPT — aziz4ai · 2026-11-07
- Open-source tool turns Garmin health data into daily LLM agent check-ins — unixterminal · 2026-11-07
- Ex-OpenAI employee leaves to build Supergroup, a super-powered group chat app — josh_wills · 2026-11-07
- AI Scribes Didn't Reduce Clinician Burden — They Just Moved It From Typing to Vigilance — mattlungrenMD · 2026-11-07