Topic
All digests tagged Information Retrieval
Suggested route
Move from earlier context through the available material to the newest signals. This route is a reading aid, not a difficulty ranking.
The fix involves 'native filtering' over a clustered vector index, which combines pre- and post-filtering to achieve 120ms P90 searches across 400 million vectors.
Read article
Kepler's architecture ensures LLMs cite numbers but never write them. By integrating deterministic code, a citation registry from 30M SEC documents, and a post-processing filter, the platform guarantees verifiable, auditable financial data.
Read article
Bridgewater's Vault processes 100M+ documents, transforming unstructured data into a first-class asset. Learn how they built for both human precision and agent scale using advanced filtering and proactive AI agents.
Read article
Exa Snapshot enables version-controlled search, allowing agents to search the web as it existed on any past date. This is critical for backtesting predictive signals and preventing model overfitting.
Read article
Learn how to build a highly accurate RAG assistant for complex financial rulebooks. We cover hybrid retrieval (BM25 + Embeddings), multi-tool agents, and streaming UX for production-grade AI.
Read article
DoorDash uses LLMs to solve semantic understanding in marketplace discovery. The core pattern is to perform expensive LLM reasoning offline and distill it into small, fast models for serving, improving retrieval and ranking.
Read article
The talk details building the document context layer for AI agents, arguing that RAG requires a three-layer platform: parsing, semantic storage, and repeatable workflows. It addresses the complexity of parsing PDFs and introduces the ParseBench benchmark.
Read article
Fixed chunking is a flawed assumption in RAG. The optimal chunk size is query-dependent. By implementing multiscale indexing and combining results using RRF, systems can achieve 20-40% higher recall.
Read article
Designing AI agents for knowledge work requires mimicking human orchestration. Agents must break down complex, ambiguous problems using specialized sub-agents and multiple search primitives to overcome performance ceilings.
Read article
Exa provides a specialized search engine for AI agents, predicting that machine searches will exceed human searches by a thousandfold by 2026. It addresses the limitations of traditional recommendation-based search.
Read article
BM25, a 30-year-old lexical scoring function, is making a comeback for Agentic Search. The change is not in the function, but in the user: the LLM agent, which can formulate highly specific queries and execute complex search trajectories.
Read article
Jean-Denis Greze argues that multi-agent systems should be reframed as search problems. The core challenge is privacy, which limits the LLM's access to all world data.
Read article
Go-to-market must be treated as an AI engineering problem. The core thesis is that GTM is fundamentally a data problem, requiring the creation of a 'live model of your world' that autonomous agents can act upon.
Read article
Discover how Multi-Vector Retrieval solves the limitations of single vector embeddings for AI agents. Learn about sparse multi-vector encoding, scaling to billion documents, and achieving sub-50ms P99 latency.
Read article
Context Engineering is the critical practice of structuring information passed to LLMs and AI agents. Learn how optimizing context, rather than just writing better prompts, improves reliability and reduces hallucinations in complex AI systems.
Read article