Enhancing Generative AI with Retrieval Augmented Generation 

Duration: 2 Days / 18 hrs

Introduction 

This 2-day, hands-on course focuses on enhancing Generative AI applications using Retrieval Augmented Generation (RAG). Participants learn how to ground large language models with enterprise data, design robust retrieval pipelines, evaluate RAG performance, and secure GenAI systems for real-world deployment. 

Objectives 

By the end of this Retrieval Augmented Generation training, participants will be able to: 

  • Understand RAG architecture, embeddings, and end-to-end workflows 
  • Build efficient indexing and retrieval pipelines using vector stores 
  • Generate accurate, context-aware responses from LLM-based systems 
  • Evaluate, optimize, and secure RAG-powered GenAI applications 

Key Takeaways 

Participants will leave with: 

  • Practical experience building RAG-based GenAI applications 
  • Clear understanding of vector databases, retrieval strategies, and prompt design 
  • Proven techniques for evaluating accuracy, cost, and ROI of RAG systems 
  • Awareness of security risks and mitigation strategies in RAG architectures 

Training Methodology (Learning by Doing) 

  • Hands-on labs throughout—every concept is implemented live 
  • Build-as-you-go approach covering indexing, retrieval, and generation 
  • Real enterprise data scenarios, not toy examples 
  • Evaluate, tune, and secure RAG systems using industry-standard techniques 

Course Outline: 

Basics of Retrieval Augmented Generation 

  • What are embeddings? 
  • RAG Phases: Indexing, Retrieval, Generation 
  • Indexing: Preparation, Chunking, Enrichment, Embeddings 
  • Working with Structured and Unstructured Documents 
  • Retrieval: Vector, Full-Text, Fusion 
  • Filtering, Reranking, Looping 
  • Generation: Providing Context to LLM-Based Systems 

Vector Stores and Indexing 

  • Vector Stores vs Other Data Stores 
  • Selecting the Right Vector Store 
  • Preparing Structured & Unstructured Data 
  • Chunking Techniques for Large Documents 
  • Enrichment and Augmentation 
  • Selecting an Embedding Model 

Retrieving Data 

  • Vector Search Algorithms & Similarity Metrics 
  • Vector, Text, and Hybrid Retrieval 
  • Multi-Query, Augmentation, Decomposition, Rewriting 
  • HyDE Technique 
  • Filtering and Reranking 
  • Retrieval Loops 

Generation Based on Context 

  • Augmenting Generation with Retrieval Results 
  • Prompting with Positive and Negative Instructions 
  • Information Extraction and Summarization 
  • Citing and Quoting Retrieved Information 
  • Prompt Chaining & Multi-Step Generation 

Evaluating RAG Systems 

  • Generative vs Predictive AI Evaluation 
  • Evaluation Metrics and Techniques 
  • RAGAS and Holistic RAG Evaluation 
  • Piecewise Evaluation 
  • Building and Selecting Evaluation Datasets 
  • Cost-Effectiveness and ROI 

Security of RAG Systems 

  • GenAI Security vs Traditional Cybersecurity 
  • Applying the CIA Triad 
  • Inversion Attacks, Data Leakage, Access-Control Mismatches 
  • Data Discovery, Poisoning, Misinformation 
  • Indirect Prompt-Injection Attacks