Course Outline: Generative AI, RAG, and Agents

Duration: 2 Months
Mode: Physical
Software Covered: LangChain, Python & Various LLMs (GPT, Gemini, Llama, Claude)

Course Overview

This course is designed to build foundational and advanced skills in Generative AI, large language models, retrieval-augmented generation (RAG) systems, and AI agents. Students will learn how to engineer prompts, implement RAG with various data sources like files, web links, and databases using LangChain, create agents capable of internet searching, and culminate in a comprehensive final project integrating these concepts for real-world applications.

Course Objectives

• Understand the fundamentals of Generative AI and its applications
• Gain practical skills in LLMs, diffusion models, and prompt engineering
• Build RAG applications using LangChain for diverse data sources including files, web content, and databases
• Develop AI agents with capabilities like internet searching

Module 1: Introduction to Generative AI

• Exploring the definition and core concepts of Generative AI
• Examining real-world applications and practical use cases in industry
• Overview of ethical considerations in Generative AI development
• Introduction to key algorithms and training methodologies
• Discussion on future trends and emerging technologies

Module 2: Introduction to LLMs & Diffusion models

• Overview of popular models like GPT, Gemini, Llama, and Claude
• Techniques for effective prompt engineering and optimization strategies
• Generating images and videos from textual descriptions using diffusion
• Comparing architectures and performance of various LLMs
• Hands-on exercises with basic model interactions and outputs

Module 3: LLM with RAG

• Understanding vector databases for efficient data storage and retrieval
• Building systems to chat interactively with your own data sources
• Creating custom GPTs tailored to specific user needs and domains
• Implementing function calling mechanisms within large language models
• Exploring integration challenges and best practices for RAG setups
• Practical examples of enhancing LLM accuracy with external knowledge

Module 4: Introduction to LangChain Framework

• Fundamentals of LangChain components and their modular structure
• Constructing simple chains and pipelines for AI workflows
• Integrating various LLMs seamlessly with the LangChain ecosystem
• Overview of memory management and state handling in chains
• Hands-on setup of development environment for LangChain projects
• Introduction to advanced features like routers and evaluators

Module 5: RAG Applications with Files

• Techniques for reading and processing PDF files with LLMs integration
• Handling Word documents in RAG systems for data extraction
• Processing Excel spreadsheets to enable querying and insights generation
• Building RAG for other file types like CSV or text documents
• Hands-on projects: Creating interactive chatbots for document-based querying

Module 6: RAG with Web Content

• Methods for loading and indexing data from specified web links
• Implementing RAG systems for retrieving web-based information efficiently
• Handling dynamic web content and real-time updates in queries
• Hands-on projects: Developing question-answering systems aware of web data

Module 7: RAG with Databases

• Integrating SQL databases with LangChain for structured data access
• Connecting NoSQL databases to enable flexible data retrieval
• Techniques for querying and retrieving insights from databases
• Hands-on projects: Building RAG applications for database-driven analytics

Module 8: Advanced AI Agents and Final Project

• Designing agents that perform autonomous internet searching tasks
• Hands-on practice building multi-tool agents for complex workflows
• Final project: Integrating RAG from files, web, databases with agent capabilities