» Tag
retrieval
10 postsBuilding an LLM Wiki: Persistent Memory for AI Agents Over Live Sources
How one team built a self-updating LLM Wiki memory layer for AI agents, tackling live source drift and comparing grep, BM25, vector, and LLM retrieval.
Agentic RAG Requires a Search Budget: Stop Endless Retrieval Loops
Agentic RAG emphasizes the importance of establishing a search budget and stop conditions while offering flexible retrieval capabilities.
Cogni: MCP Memory for LLMs Without LLM in the Retrieval Path
Cogni enhances LLMs with connected memory, enabling efficient reasoning and retrieval without relying on LLMs during the process.
CommitBrief — AI code reviews, right in your terminal
A provider-agnostic, local-first CLI that reviews your staged changes, a historic range, or a whole GitHub pull request. Zero telemetry, no server. Free and open source.
commitbrief.comBuilding a Hybrid RAG API: The LLM Was the Easy Part
Challenges in building a hybrid RAG API and its importance for engineers. Details on PDF processing and question answering.
Better Embeddings Won’t Fix Missing Provenance in RAG
Explore how missing provenance can lead to inaccuracies in retrieval systems, despite better embeddings.
Vector search isn't the hard part; deciding what to search is
Exploring the importance of metadata and pre-search decisions in AI retrieval systems.
Coalent: An LLM Answer Cache That Invalidates When Source Changes
Coalent offers an LLM answer cache that automatically invalidates when sources change.
Multi-Vector (Late Interaction) Embedding Models with Sentence Transformers
The v6.0 update of Sentence Transformers introduces MultiVectorEncoder for enhanced late interaction retrieval.
When GraphRAG Is Overkill: A Break-Even Test for RAG and Graphs
Explore the importance of question classification in evaluating GraphRAG's effectiveness and cost. Discover the balance between point lookups and complex relationships.
My D&D Campaign Notes Accidentally Matched Google's New RAG Spec
My D&D campaign notes system accidentally matched Google's Open Knowledge Format. How did this impact my querying processes?