Mastering Retrieval-Augmented Generation

Vector databases, hybrid search, cross-encoder rerankers, and production chunking strategies.

One-time purchase

£29.00

Instant Access

I consent to immediate access to this digital content and acknowledge that under the UK Consumer Contracts Regulations 2013, my 14-day statutory right of cancellation is waived upon purchase.

Encrypted payment processing via Stripe Checkout

DRM-free PDF e-book delivered immediately after payment

Purchase receipt emailed automatically

About this Book

Naive RAG fails in production. This comprehensive manual details the complete pipeline required to build high-accuracy question-answering systems over millions of internal documents: semantic chunking, reciprocal rank fusion (RRF), parent-document retrievers, contextual compression, and automated evaluation with Ragas.

Target Audience: Software engineers and ML practitioners building enterprise search and knowledge bases.

What You Will Learn & Build

Master chunking heuristics: semantic boundaries, markdown headers, and recursive splitting

Combine dense vector search with sparse BM25 keyword matching via Reciprocal Rank Fusion

Deploy cross-encoder reranking models to boost top-3 hit rate by over 40%

Detect and eliminate retrieval hallucinations before sending context to the generator

1. Why Naive RAG Collapses Under Real WorkloadsSection 1

2. Document Parsing, OCR & Clean Extraction PipelinesSection 2

3. Semantic vs Recursive Chunking HeuristicsSection 3

4. Vector Embeddings: Dimensionality, Distance Metrics & NormalizationSection 4

5. Hybrid Search: Dense Vectors + BM25 with Reciprocal Rank FusionSection 5

6. Cross-Encoder Rerankers & Context CompressionSection 6

7. Benchmarking RAG Quality with Ragas and TruLensSection 7

Publication Specifications

Mastering Retrieval-Augmented Generation

About this Book

What You Will Learn & Build

Table of Contents