Applied Generative AI: Building RAG Systems for Enterprise
Scheduled sessions
Beyond the chat box: Move from simple LLM wrappers to robust, enterprise-grade Generative AI applications using your company's private data.
Master the RAG (Retrieval-Augmented Generation) architecture. Learn how to ingest complex documents, apply advanced chunking strategies, and connect semantic search with Large Language Models.
Frameworks & Tooling: Get hands-on with industry-standard orchestration frameworks like LangChain and LlamaIndex to build complex AI pipelines.
Production Focus: ~70% hands-on labs focused not just on building RAG, but on evaluating it. Learn techniques to mitigate hallucinations, measure recall/precision (using RAGAS), and handle prompt injection.
Who it’s for: Software Engineers, AI Developers, and Data Scientists tasked with delivering reliable Generative AI solutions.
Skills You Will Learn
Curriculum
LLM Fundamentals & API Integration
- Understanding LLMs: Tokens, context windows, temperature, and generation limits
- Interacting with LLM APIs (OpenAI) and local models (Ollama / HuggingFace)
- Advanced Prompt Engineering: Few-shot prompting, Chain-of-Thought (CoT), and formatting instructions
- Lab: Building a structured data extractor using function calling / JSON mode
The RAG Architecture & Data Ingestion
- What is Retrieval-Augmented Generation (RAG) and why is it necessary?
- Document Loaders: Ingesting PDFs, Confluence pages, and Markdown files
- Advanced Chunking Strategies: Recursive character splitting, semantic chunking, and handling overlapping
- Lab: Building an automated data ingestion pipeline into a vector store
Orchestration Frameworks: LangChain & LlamaIndex
- Introduction to LangChain: Chains, Prompts, and Output Parsers
- LlamaIndex basics: Nodes, Indices, and Query Engines
- Connecting retrievers to LLMs: Stuffing, Map-Reduce, and Refine document chains
- Lab: Building a conversational QA bot over a technical documentation repository
Evaluating and Productionizing RAG
- The problem of hallucinations and how to mitigate them
- RAG Evaluation: Measuring context precision, recall, and answer relevancy (using RAGAS/TruLens)
- Query transformations: Query re-writing, sub-queries, and hybrid search
- Lab: Implementing an evaluation pipeline and tuning chunk sizes to improve mAP (Mean Average Precision)
Optional modules
Optional — Advanced Retrieval Techniques
- Implementing Re-ranking (Cross-Encoders) to improve top-K results
- Parent-Document Retrieval and small-to-big retrieval patterns
- Self-RAG: Teaching the LLM to critique its own retrieved context
Course Day Structure
- Part 1: 09:00–10:30
- Break: 10:30–10:45
- Part 2: 10:45–12:15
- Lunch break: 12:15–13:15
- Part 3: 13:15–15:15
- Break: 15:15–15:30
- Part 4: 15:30–17:30