Build Intelligent AI Apps with Retrieval-Augmented Generation (RAG)

• Demystify Retrieval-Augmented Generation and its components
• Enable real-world implementation with popular AI toolkits
• Teach prompt injection prevention, grounding, and retrieval optimization
• Stimulate innovation in building knowledge-enhanced LLM apps
• Guide participants to build and deploy a fully functional AI app with RAG

Day 1: Introduction & Fundamentals of RAG

• Overview of Language Models (LLMs) & Limitations
• Why RAG? – Bridging LLMs with External Knowledge
• Anatomy of a RAG System: Retriever + Generator
• Tools: FAISS, ChromaDB, LangChain, HuggingFace, OpenAI

Hands-on:

• Build a vector store using FAISS/Chroma
• Embed documents using SentenceTransformers or OpenAI Embeddings

Day 2: Building the RAG Pipeline

• Types of Retrievers: Dense vs Sparse
• Query Rewriting, Chunking & Indexing
• LangChain vs Haystack vs Custom RAG Pipelines
• Prompt Engineering in RAG

Hands-on:

• Build a RAG pipeline using LangChain
• Query documents and generate contextual answers using GPT

Day 3: Advanced Features & Deployment

• Evaluation Metrics for RAG Systems (EM, BLEU, Recall@k)
• Caching, Logging, and Observability
• Securing APIs & Cost Management
• Open-source RAG use cases (e.g., DocChat, Chat with PDF, Legal AI, Academic Copilots)

Hands-on:

• Deploy a working RAG chatbot on Gradio
• Custom PDF/CSV Q&A assistant (bring-your-own-data)

• Master RAG pipeline design and deployment
• Learn to embed documents and retrieve relevant context dynamically
• Combine search and generation seamlessly for fact-rich AI applications
• Build an intelligent AI assistant that can reason over documents or databases
• Gain hands-on experience with LangChain, FAISS, Pinecone, and Hugging Face
• Receive international workshop certification and reusable project templates

• AI/ML developers and data scientists
• Backend or full-stack developers building AI integrations
• NLP researchers and LLM enthusiasts
• LegalTech, HealthTech, and EdTech innovators
• Students with a working knowledge of Python and NLP