Category
Grounding AI in your data
A language model only knows what it was trained on. The retrieval layer fixes that: it finds the right pieces of your data at the right moment and puts them in front of the model. These entries cover how that works, from embeddings and vector databases to chunking, reranking, and full RAG pipelines.
Entries in this category
- Chunking Splitting documents into smaller pieces so each one can be embedded and retrieved on its own. Read the entry
- Data pipelines for AI The ingestion, cleaning, chunking, and embedding steps that turn raw data into a searchable index. Read the entry
- Embedding models A model that converts text into a vector of numbers so similar meanings land near each other. Read the entry
- Hybrid search Running keyword and vector search together and merging results, so exact terms and meaning both count. Read the entry
- Knowledge graph A network of entities and their relationships that lets AI systems follow facts instead of guessing at them. Read the entry
- pgvector A Postgres extension that stores embeddings and runs similarity search inside the database you already have. Read the entry
- RAG (retrieval augmented generation) Fetching relevant documents at query time so a language model can answer from your data instead of memory alone. Read the entry
- Reranking A second scoring pass that reorders retrieved candidates so the most relevant results come first. Read the entry
- Semantic search Search that ranks results by meaning, using embeddings, so matches do not require the exact words. Read the entry
- Vector database A database that stores embeddings and finds nearest neighbors fast, powering semantic search and RAG. Read the entry
Where this layer fits
Retrieval sits between your data and the model. Upstream, core concepts like embeddings and context windows explain the raw materials. Downstream, agents decide when to retrieve, and evals tell you whether the answers actually got better.
Not sure where to start? RAG is the entry that ties this whole category together. Or browse everything at once in the A-Z index.