Method

BM25

Classic lexical ranking function for keyword search.

29 cases from 21 companies apply BM25. 2 were checked in an online A/B test; 5 report results in production.

Typical tasks: Build an assistant over internal docs, Rank search results, Evaluate models, Retrieve and rank from millions of items and 7 more. Companies: Avito, Sber, VTB, X5 Tech, Algolia, Databricks and 15 more. Often combined with: Hybrid search, RAG, Cross-encoder reranking, LLM-as-a-judge, Reciprocal Rank Fusion and 82 more.

Effective document management and retrieval for generative AIbanking2026-06
AI Agents for the Frontline: 7-Eleven's GenAI Maintenance Assistantretail2026-06
How a 4B Model Beat GPT at Agentic Retrieval (And How to Train One on a Neon Postgres Corpus)general2026-06
Multimodal Soup: Turning a Heap of Ad Banners into Meaningful Clustersadtech2026-05
From an LLM Portal to a Factory of Internal Agentsclassifieds2026-05
An AI Tutor and Methods for Evaluating Itedtech2026-05
Diagnosing LLM-Based Ranking Models in Cold-Start Recommender Systems: Coverage, Diversity and Practical Fixesmedia-streaming2026-05
The Search of the Futuregeneral2026-05
Building a Battle-Tested Open-Source Agentic RAGgeneral2026-05
RAG for Complex Situational Questionsbanking2026-05
An Automated Request Handling Service for SMBsbanking2026-05
Winner's Solution Walkthrough: 1st Place (AvitoTech ML Cup 2026)classifieds2026-05
Building RAG Systems That Work Accurately: From Prototype to Thousands of Users a Dayretail2026-05
Modernizing the Facebook Groups Search to Unlock the Power of Community Knowledgesocial2026-04
Hybrid Search — Where Keywords Meet Vectors, Enabling Classifieds Discoveryclassifieds2025-09
Unraveling Rovo search relevancesaas-b2b2025-08
How LLM Automates Work at Avitoclassifieds2025-05
Graphs and Agents in Action: Building a CoPilot for All Sber Branchesbanking2025-05
Make Retrieval Smarter: How We Built Smart Search over a Document Basegeneral2025-05
Retrieval-Augmented Fill-in-the-Middle: An Architecture for Generating Structured Documents with LLMsgeneral2025-05
In IDE Code Retriever—general2024-11
Multimodality and RAG, or How to Sit on Two Chairsgeneral2024-11
Machine Learning Service from Scratch, or Creating Frankensteingeneral2024-11
Advanced application of GigaChat at Rabota.ru. Recommendations and auto-moderationclassifieds2024-05
How We Got RAG Running in Productiongeneral2024-05
RAG Tricks: Practical Steps for Corporate Document RAGgeneral2024-05
Why weights are often counterproductive in rankinge-commerce2023-12
Comparing Algolia and Elasticsearch For Consumer-Grade Search Part 2: Relevance Isn't Luckgeneral2023-08
3 Changes to Expand DoorDash's Product Search Beyond Deliveryfood-delivery2022-05