GenAI PMDoug Turnbull
Search and retrieval expert mentioned for introducing pseudo-relevance feedback. He explains how early retrieval results can be used to refine queries.
Key Highlights
- Doug Turnbull is a recurring expert voice on search, retrieval, and relevance engineering for AI products.
- He is specifically noted for explaining pseudo-relevance feedback and how early results can improve later retrieval.
- His writing gives practical guidance on choosing between search engines, vector databases, and hybrid architectures.
- He also contributes useful frameworks for evaluating search quality with LLM-based relevance methods.
At a glance
- Role
- Search and retrieval expert
Doug Turnbull
Overview
Doug Turnbull is a search and retrieval expert frequently cited for practical, technically grounded thinking on information retrieval, search relevance, and RAG system design. In the newsletter coverage, he appears as a consistent source of insight on topics such as BM25, pseudo-relevance feedback, retrieval engine selection, late interaction models, and evaluation methodology for search quality.For AI Product Managers, Turnbull matters because his work sits at the intersection of classic search, modern retrieval, and production AI systems. His writing helps PMs make better decisions about when to use lexical search versus vector search, how to evaluate retrieval quality, how to refine queries using early results, and how to choose infrastructure based on real trade-offs rather than market hype.
Key Developments
- 2026-02-03 — Discussed LLM search relevance evaluation, emphasizing the importance of checking both directions in pairwise evaluation to avoid misleading judgments.
- 2026-03-07 — Explored whether BM25 can be interpreted probabilistically, introducing a Bayesian framing and implications for calibrating hybrid search systems.
- 2026-03-11 — Argued in The tests are the code now that AI-assisted coding increases the importance of tests as the core mechanism for preserving software quality.
- 2026-03-21 — Published guidance on how to choose a retrieval engine, comparing search engines and vector databases such as Elasticsearch, OpenSearch, Solr, Vespa, Pinecone, Turbopuffer, and Weaviate.
- 2026-03-24 — Covered why tiny late interaction models can win, referencing a LightOn demonstration with developer Antoine Chaffin and discussing implications for retrieval architecture.
- 2026-04-07 — Argued that a RAG-style system can technically be built with grep if a team is willing to do substantial engineering work, while stressing the practical difficulty of that path.
- 2026-04-14 — Explained pseudo-relevance feedback, showing how initial ranked results from BM25 or other retrieval methods can be reused to refine queries and improve downstream relevance.
Relevance to AI PMs
- Design better retrieval stacks: Turnbull’s work helps PMs decide when to use lexical retrieval, vector retrieval, or hybrid-search, and how to reason about trade-offs in cost, latency, controllability, and relevance.
- Improve search quality evaluation: His writing on llm-search-relevance-eval and llm-pairwise-evaluation is useful for PMs building eval loops, especially when testing rankers, RAG retrieval, or query rewriting systems.
- Make more informed platform choices: His comparisons across engines and databases give PMs a practical framework for selecting infrastructure like Elasticsearch, OpenSearch, Solr, Vespa, Pinecone, Turbopuffer, and Weaviate based on use case rather than vendor narratives.
Related
- pseudo-relevance-feedback — A core concept Turnbull is specifically mentioned for explaining; it uses top retrieval results as implicit feedback to refine subsequent queries.
- bm25 — Central to several of his discussions, including probabilistic interpretations and its role as a baseline for retrieval.
- hybrid-search — Connected through his analysis of combining lexical and probabilistic or vector-based retrieval signals.
- rag — His commentary on building retrieval systems, including the provocative grep framing, directly informs RAG architecture choices.
- elasticsearch, opensearch, solr, vespa — Traditional and modern search engines he discusses when helping teams choose retrieval infrastructure.
- pinecone, turbopuffer, weaviate — Vector-oriented platforms included in his retrieval engine comparison framework.
- lighton, antoine-chaffin — Referenced in his discussion of tiny late interaction models and retrieval innovation.
- tests-as-the-code — Related through his argument that testing becomes the primary quality artifact in AI-assisted software development.
- llm-search-relevance-eval, llm-pairwise-evaluation — Strongly connected to his views on evaluating search relevance with LLM-based methods.
Frequently asked questions
What is Doug Turnbull?
Search and retrieval expert mentioned for introducing pseudo-relevance feedback. He explains how early retrieval results can be used to refine queries.
What should AI product managers know about Doug Turnbull?
Doug Turnbull is a recurring expert voice on search, retrieval, and relevance engineering for AI products. He is specifically noted for explaining pseudo-relevance feedback and how early results can improve later retrieval. His writing gives practical guidance on choosing between search engines, vector databases, and hybrid architectures.
Explore this topic
Newsletter Mentions (7)
“Doug Turnbull What is psuedo-relevance feedback? - Introduces pseudo-relevance feedback: after an initial BM25 or ranked retrieval, the returned results provide implicit information that can be used to refine queries or improve subsequent retrieval.”
#7 📝 Doug Turnbull What is psuedo-relevance feedback? - Introduces pseudo-relevance feedback: after an initial BM25 or ranked retrieval, the returned results provide implicit information that can be used to refine queries or improve subsequent retrieval. The post outlines how to leverage those initial results as a source of feedback to boost relevance.
“#11 📝 Doug Turnbull Is grep all you need for RAG? - Doug argues that with enough engineering effort you can build a RAG-style search system using only grep, but cautions that this approach is difficult and not for the faint of heart.”
#11 📝 Doug Turnbull Is grep all you need for RAG? - Doug argues that with enough engineering effort you can build a RAG-style search system using only grep, but cautions that this approach is difficult and not for the faint of heart.
“Doug Turnbull Why tiny late interaction models win - Discusses the recent prominence of late interaction models, highlighting a LightOn demonstration (with developer Antoine Chaffin) using a 150M model and the implications for retrieval and interaction approaches.”
#12 📝 Doug Turnbull Why tiny late interaction models win - Discusses the recent prominence of late interaction models, highlighting a LightOn demonstration (with developer Antoine Chaffin) using a 150M model and the implications for retrieval and interaction approaches.
“Doug Turnbull How to actually choose a retrieval engine - Explains how teams should choose a retrieval engine by comparing vector databases and search engines, and considering trade-offs between options like Elasticsearch, OpenSearch, Solr, Vespa, Pinecone, Turbopuffer, and Weaviate.”
#5 📝 Doug Turnbull How to actually choose a retrieval engine - Explains how teams should choose a retrieval engine by comparing vector databases and search engines, and considering trade-offs between options like Elasticsearch, OpenSearch, Solr, Vespa, Pinecone, Turbopuffer, and Weaviate. Emphasizes practical selection criteria beyond hype.
“#15 📝 Doug Turnbull The tests are the code now - Argues that with AI-assisted coding, tests become the most important artifact for maintaining code quality.”
Doug Turnbull is cited for a piece arguing that AI-assisted coding elevates the importance of tests. The newsletter uses him to support a broader software quality point.
“#20 📝 Doug Turnbull Can BM25 be a probability? - Explores the relationship between BM25 scores framed as odds versus probabilities and introduces a Bayesian view of BM25.”
GenAI PM Daily March 07, 2026 GenAI PM Daily 🎧 Listen to this brief 3 min listen Today's top 25 insights for PM Builders, ranked by relevance from LinkedIn, YouTube, X, and Blogs. #20 📝 Doug Turnbull Can BM25 be a probability? - Explores the relationship between BM25 scores framed as odds versus probabilities and introduces a Bayesian view of BM25. Discusses implications for calibrating hybrid search systems when combining lexical and probabilistic signals.
“Check twice, cut once with LLM search relevance eval - Highlights the importance of checking both directions in LLM pairwise evaluation of search relevance.”
GenAI PM Daily February 03, 2026 GenAI PM Daily Today's top 10 insights for PM Builders, ranked by relevance from Blogs, X, YouTube, and LinkedIn. OpenAI Launches Codex App 📝 OpenAI News Introducing the Codex app - OpenAI has launched the Codex app, enhancing user interaction with AI. Read more → 𝕏 claire vo 🖤 @clairevo Claire overhauled Maplewood’s architecture by migrating to Inngest workflows and persisting stories/actions in NeonDB, added infinite scroll for event feeds, and squashed an auto-scroll bug. Read more → 📝 Doug Turnbull Check twice, cut once with LLM search relevance eval - Highlights the importance of checking both directions in LLM pairwise evaluation of search relevance.
Related
RAG is a retrieval-based pattern that injects external context into prompts to improve model responses. The newsletter presents it as often outperforming fine-tuning for practical product work.
A lexical retrieval ranking function used here to select relevant tool definitions. In PM tooling, it helps improve retrieval accuracy and reduce context-window bloat.
Elasticsearch is referenced in the context of hybrid search and kNN query behavior in practice.
Stay updated on Doug Turnbull
Get curated AI PM insights delivered daily — covering this and 1,000+ other sources.
Subscribe Free