This course contains the use of artificial intelligence.

Move beyond basic Retrieval-Augmented Generation demos and learn how to design, build, evaluate, secure, and operate production-ready enterprise RAG systems.

In this hands-on course, you will build an Enterprise Knowledge Intelligence Platform that evolves throughout the curriculum. You will begin with a baseline RAG application and progressively add advanced ingestion, chunking, retrieval, query enhancement, adaptive workflows, graph-based retrieval, multimodal document processing, evaluation, observability, security, and deployment capabilities.

You will learn how to process PDFs, HTML, Office documents, spreadsheets, tables, and scanned files. You will compare fixed, recursive, structure-aware, semantic, hierarchical, and parent-child chunking strategies while preserving metadata, document versions, permissions, and source lineage.

You will implement hybrid retrieval using keyword search, dense embeddings, metadata filters, Reciprocal Rank Fusion, and CPU-compatible cross-encoder re-ranking. You will improve retrieval relevance with query rewriting, multi-query generation, query decomposition, Hypothetical Document Embeddings, intent-preservation checks, and retrieval routing.

The course also covers Self-RAG and Corrective RAG patterns. You will build bounded workflows that grade retrieved evidence, correct failed retrieval, verify generated answers, enforce retry limits, and return grounded no-answer responses when reliable evidence is unavailable.

You will create a CPU-friendly GraphRAG pipeline for entity and relationship modeling, entity resolution, provenance, graph traversal, multi-hop retrieval, and hybrid graph-document search. You will also build a CPU-only multimodal retrieval workflow for OCR, layout-aware parsing, structured tables, figures, page-level search, and visual citations.

To prepare the application for production, you will create golden evaluation datasets and measure Recall@K, Precision@K, MRR, nDCG, grounding, faithfulness, citation quality, latency, and workload. You will add local tracing, caching, context optimization, cost estimation, ACL-aware retrieval, tenant isolation, prompt-injection defenses, deletion workflows, CI/CD quality gates, and containerized deployment.

All mandatory demonstrations and Hands on Labs run locally with open-source tools. No paid AI API, cloud account, managed database, or dedicated GPU is required.

This course is designed for AI engineers, machine learning engineers, software developers, data engineers, platform engineers, solutions architects, MLOps professionals, and technical leads who already understand basic RAG concepts and want to build reliable enterprise AI systems.