RAGEngineering Hub
Menu

Knowledge map

How the concepts connect

Each column is a Part. A line means the page on the right builds on the page on the left. Scroll sideways to see all 8 Parts.

PART 1PART 2PART 3PART 4PART 5PART 6PART 7PART 8RAG-01Why models need outside knowl…Why models need outside knowledge, and what hallucination meansRAG-02What RAG is: anatomy of a req…What RAG is: anatomy of a requestRAG-03Tokens and context windows fo…Tokens and context windows for RAGRAG-04RAG vs fine tuning vs long co…RAG vs fine tuning vs long context and prompt cachingRAG-05The RAG lifecycle and your fi…The RAG lifecycle and your first RAG in miniatureRAG-06Use cases and the quality, la…Use cases and the quality, latency and cost triangleRAG-07Content sources and file form…Content sources and file formatsRAG-08Parsing, layout and OCRParsing, layout and OCRRAG-09Tables, images and mixed cont…Tables, images and mixed contentRAG-10Cleaning, normalization and q…Cleaning, normalization and quality gatesRAG-11Metadata, document identity a…Metadata, document identity and versioningRAG-12Updates, deletes, reindexing …Updates, deletes, reindexing and ingestion pipelinesRAG-13Why we chunk, and the main ch…Why we chunk, and the main chunking strategiesRAG-14Structure aware chunking and …Structure aware chunking and tablesRAG-15Chunk size, overlap and token…Chunk size, overlap and token based sizingRAG-16Contextual chunking and late …Contextual chunking and late chunkingRAG-17Parent child and multi repres…Parent child and multi representation indexingRAG-18Chunk metadata, provenance an…Chunk metadata, provenance and the chunking debugging playbookRAG-19Embeddings and similarity fro…Embeddings and similarity from first principlesRAG-20What embeddings cannot doWhat embeddings cannot doRAG-21Choosing an embedding model: …Choosing an embedding model: size, cost, languageRAG-22Vector databases, ANN indexes…Vector databases, ANN indexes and choosing a storeRAG-23Score calibration and thresho…Score calibration and thresholdsRAG-24Late interaction and multivec…Late interaction and multivector retrievalRAG-25Embedding model migration and…Embedding model migration and index rebuildsRAG-26Query understanding, rewritin…Query understanding, rewriting and expansionRAG-27Dense retrieval and keyword s…Dense retrieval and keyword search with BM25RAG-28Hybrid retrieval and rank fus…Hybrid retrieval and rank fusionRAG-29topK, no good match, and meta…topK, no good match, and metadata filteringRAG-30Routing, diversity, deduplica…Routing, diversity, deduplication and cachingRAG-31Reranking: two stage, cross e…Reranking: two stage, cross encoder and LLM rerankersRAG-32Retrieval failure analysis pl…Retrieval failure analysis playbookRAG-33Context selection, packing an…Context selection, packing and citationsRAG-34Grounding prompts and abstent…Grounding prompts and abstentionRAG-35Lost in the middle and contex…Lost in the middle and context managementRAG-36Conflicts, staleness and fres…Conflicts, staleness and freshness rulesRAG-37Hallucination checks and answ…Hallucination checks and answer verificationRAG-38Conversational RAG, answer UX…Conversational RAG, answer UX and RAG APIsRAG-39Why measure, and what to meas…Why measure, and what to measureRAG-40Building a golden datasetBuilding a golden datasetRAG-41Retrieval metrics: precision,…Retrieval metrics: precision, recall, MRR, nDCGRAG-42Faithfulness and answer corre…Faithfulness and answer correctnessRAG-43LLM as judge, and judging the…LLM as judge, and judging the judgeRAG-44Offline and online evals, reg…Offline and online evals, regression tests and slicesRAG-45Prompt injection and poisoned…Prompt injection and poisoned documentsRAG-46Access control, privacy and g…Access control, privacy and governanceRAG-47Index lifecycle operationsIndex lifecycle operationsRAG-48Observability and tracingObservability and tracingRAG-49Latency, cost, caching and mo…Latency, cost, caching and model routingRAG-50Reliability, degraded modes a…Reliability, degraded modes and launch readinessRAG-51Scaling, multi tenancy and en…Scaling, multi tenancy and enterprise architectureRAG-52Choosing an architecture: nai…Choosing an architecture: naive, advanced, agentic, graphRAG-53Agentic RAG and RAG as an MCP…Agentic RAG and RAG as an MCP toolRAG-54GraphRAG, structured data and…GraphRAG, structured data and multimodal RAG