# Search and retrieval A pack of 10 talks from the AI Engineer YouTube channel, in the order to watch them. 4h 16m of video. Page: https://aietalks.com/packs/search-and-retrieval A fluent RAG answer can still be wrong because the right passage never reached the model. The retrieval engine comes first: inverted indexes, BM25, dense vectors and the ranking signals that decide what survives. One practical rule carries through the harder cases: inspect the bad queries before adding another layer. Filters, rerankers and query decomposition earn their place that way. Writer's graph work shows why repeated enterprise language can confuse vector similarity, while a FIFA corpus makes the case for text-to-SQL when the question asks for counts or comparisons across the whole collection. Agent search only helps once that retrieval layer works. Its tests catch web access that silently failed and require current evidence, so a bad score can suggest an actual repair. ## This pack is for you if - Your RAG system returns plausible passages but still misses the evidence needed to answer. - You are deciding between lexical, vector, graph and structured retrieval for a real corpus. - Your agent claims it searched, yet you cannot tell what it found or whether the sources were current. ## The talks, in order ### 1. Information Retrieval from the Ground Up Philipp Krenn, Elastic | 1:48:07 | AI Engineer World's Fair 2025 Video: https://www.youtube.com/watch?v=4Xe_iMYxBQc Summary: https://aietalks.com/talks/information-retrieval-from-the-ground-up.md Why start here: Krenn strips retrieval back to the machinery beneath the current RAG vocabulary: analysis, inverted indexes, BM25, dense vectors and ranking. The point is not to crown one method, but to understand the signals each can recover. Liu then shows how those signals appear in a modern AI search system. ### 2. The State of AI-Powered Search and Retrieval Frank Liu, MongoDB | 12:35 | AI Engineer World's Fair 2025 Video: https://www.youtube.com/watch?v=pIPtpBZ6TKk Summary: https://aietalks.com/talks/the-state-of-ai-powered-search-and-retrieval.md Why second: Krenn supplies the foundations; Liu assembles them into an application stack with semantic search, metadata filters, query decomposition and multimodal inputs. His overview makes the menu visible. Karam follows with the harder rule for choosing from it: add a technique only when a failing query earns it. ### 3. Layering Every Technique in RAG, One Query at a Time David Karam, Pi Labs | 20:22 | AI Engineer World's Fair 2025 Video: https://www.youtube.com/watch?v=w9u11ioHGA0 Summary: https://aietalks.com/talks/layering-every-technique-in-rag-one-query-at-a-time.md Why here: Liu describes the available layers. Karam orders them by evidence, starting with the simplest system and reading its losses before adding BM25, embeddings, rerankers, domain signals or query fan-out. Fernandes then turns that method into a close-up of one naive RAG stack failing and being repaired. ### 4. The RAG Stack We Landed On After 37 Fails Jonathan Fernandes | 18:52 | AI Engineer World's Fair 2025 Video: https://www.youtube.com/watch?v=2CXn-CByNoo Summary: https://aietalks.com/talks/the-rag-stack-we-landed-on-after-37-fails.md Why here: Karam gives a decision process; Fernandes shows the debugging experience it produces. Changing embedding models does not fix bad context, while a cross-encoder reranker finally moves the right passage into place. Ma follows by widening that local repair into the economic and architectural choices of an enterprise retrieval stack. ### 5. RAG in 2025: State of the Art and the Road Forward Tengyu Ma, MongoDB | 18:48 | AI Engineer World's Fair 2025 Video: https://www.youtube.com/watch?v=W_CYk2ogcDI Summary: https://aietalks.com/talks/rag-in-2025-state-of-the-art-and-the-road-forward.md Why here: Fernandes improves one stack through traces and evaluation. Ma asks why an enterprise should retrieve at all, then covers access control, vector compression, hybrid search and multimodal documents. His warning that performance varies sharply by corpus sets up Julien's case, where dense repetitive documents make vectors unreliable. ### 6. When Vectors Break Down: Graph-Based RAG for Dense Enterprise Knowledge Sam Julien, Writer | 15:47 | AI Engineer World's Fair 2025 Video: https://www.youtube.com/watch?v=XlAIgmi_Vow Summary: https://aietalks.com/talks/when-vectors-break-down-graph-based-rag-for-dense-enterprise-knowledge.md Why here: Ma presents graph retrieval as one tool among several. Julien gives the failure that justifies it: enterprise documents reuse the same language while the answer depends on relationships spread across pages. Belfer and Granot then push the same challenge further, asking when retrieval should give way to a structured query. ### 7. RAG Evaluation Is Broken! Here's Why (And How to Fix It) Yuval Belfer & Niv Granot, A21 Labs | 10:58 | AI Engineer World's Fair 2025 Video: https://www.youtube.com/watch?v=Ywl4LsvHKzU Summary: https://aietalks.com/talks/rag-evaluation-is-broken-heres-why-and-how-to-fix-it.md Why here: Julien preserves relationships with graph-derived retrieval. Belfer and Granot show a different boundary: top-k chunks are a poor tool for counts, maxima and corpus-wide comparisons. Their text-to-SQL alternative makes the representation match the question. Lichtenberg follows by returning to unstructured evidence and teaching an agent to choose among retrieval tools. ### 8. How We Taught Agents to Use Good Retrieval Hanna Lichtenberg, Mixedbread AI | 14:28 | AI Engineer World's Fair 2026 Video: https://www.youtube.com/watch?v=1IdzkRVmWAA Summary: https://aietalks.com/talks/how-we-taught-agents-to-use-good-retrieval.md Why here: Belfer and Granot choose the interface that fits the data. Lichtenberg lets an agent make that choice among overview search, semantic search, metadata filters and grep, then trains it to describe evidence needs instead of guessing keywords. Levi follows with a prior condition this work assumes: the agent must reach the source at all. ### 9. Your Agent's Biggest Lie: "I Searched the Web" Rafael Levi, Bright Data | 15:49 | AI Engineer Europe 2026 Video: https://www.youtube.com/watch?v=btxGmN8RvNU Summary: https://aietalks.com/talks/your-agents-biggest-lie-i-searched-the-web.md Why here: Lichtenberg improves how an agent searches a corpus. Levi shows why a well-formed query can still produce fiction on the public web: the request hits a CAPTCHA, an empty page or stale training data, and the agent answers anyway. Neagu, Emery and Sher close the pack by making access, evidence and answer quality measurable. ### 10. Evaluating AI Search: A Practical Framework for Augmented AI Systems Julia Neagu, Deanna Emery & Mara Sher, Quotient AI | 20:33 | AI Engineer World's Fair 2025 Video: https://www.youtube.com/watch?v=wRJD0inpmjU Summary: https://aietalks.com/talks/evaluating-ai-search-a-practical-framework-for-augmented-ai-systems.md Why end here: Levi exposes the invisible failure of web retrieval. Neagu, Emery and Sher build the operating discipline around it: dynamic test sets for changing facts, separate measures for completeness, document relevance and hallucination, and results that point toward a repair. It ends the pack where production search must end, with evidence strong enough to change the system.