Build Production RAG Systems
Ship document Q&A and support bots that customers trust. This course walks you through the full production RAG stack—from ingestion and chunking through embeddings, hybrid retrieval, pipeline assembly, and deployment with eval and cost controls. Every module ends with a hands-on lab you can drop into your own codebase. **What you'll build:** - A chunking pipeline with overlap and metadata - Vector search with cosine similarity - Hybrid retrieval with reranking and MRR evaluation - A complete context assembly layer for LLM prompts - Production monitoring, eval harnesses, and cost dashboards Built for engineers who need RAG that works on real PDFs, wikis, and ticket archives—not toy demos.
Instructor: Dr. Sarah Chen
Course Curriculum
Module 1: Document Ingestion & Chunking
By the end of this module you'll build a production chunking pipeline with overlap, metadata extraction, and idempotent ingestion.
- Why Ingestion Makes or Breaks RAG
- Chunking Strategies & Metadata Design
- Parsing Pipelines for PDF, HTML & Wikis
- Idempotency, CDC & Ingestion at Scale
- Lab: Sliding-Window ChunkerLab
- Module 1: Ingestion & ChunkingQuiz
Module 2: Embeddings & Vector Search
By the end of this module you'll implement cosine similarity search and understand how to scale vector retrieval to production indexes.
- Embeddings: From Text to Geometry
- Vector Indexes & Similarity Search at Scale
- Embedding Model Benchmarking
- Multi-Tenant Vector Storage & Normalization
- Lab: Cosine Similarity SearchLab
- Module 2: Embeddings & Vector SearchQuiz
Module 3: Retrieval Quality & Hybrid Search
By the end of this module you'll implement MRR evaluation and design hybrid retrieval pipelines that beat pure vector search.
- Reranking, Hybrid Search & When Vectors Fail
- Evaluating Retrieval: MRR, Recall@K, and Golden Sets
- Query Transformation & Retrieval Tuning
- Filtered Search, ACL & Domain Retrieval
- Lab: Mean Reciprocal Rank (MRR)Lab
- Module 3: Retrieval QualityQuiz
Module 4: Building the RAG Pipeline
By the end of this module you'll build a context assembly layer that turns retrieved chunks into reliable LLM prompts.
- End-to-End RAG Architecture
- Context Assembly, Citations & Faithfulness
- Multi-Turn RAG & Query Rewriting
- Caching, Streaming & Latency Optimization
- Lab: Context AssemblerLab
- Module 4: RAG PipelineQuiz
Module 5: Production Deployment & Operations
By the end of this module you'll design eval harnesses, monitoring dashboards, and cost controls for production RAG.
- Eval Pipelines: Retrieval to Answer Quality
- Monitoring, Cost Control & Incident Response
- RAG Security & Document Poisoning
- Blue-Green Indexing & Re-Index Strategy
- Lab: RAG Cost EstimatorLab
- Module 5: Production OperationsQuiz
$94.99
Lifetime access · Certificate included
- Hands-on coding exercises
- Module knowledge check quizzes
- Downloadable resources
- Progress tracking dashboard