• What is Generative AI?
• Anatomy of LLMs (briefly covering transformer, GPT, and BERT)
• Introduction to Prompt Engineering
  •  Few-shot, Zero-shot, Chain-of-Thought
  • Output formatting (JSON, tables, etc.)
  • Prompt templates in LangChain
• Hands-on: Create prompt templates for classification and summarization

• Why RAG? Overcoming LLM context limitations
• Core components: Embeddings, Vector Stores, Chunking, Retriever
• Implementing RAG with LangChain
• Chunking strategies and embedding models
• Hands-on: Build a basic RAG system with PDF/text data

• LangChain architecture: Chains, Tools, Agents
• Memory, PromptTemplates, LLMChains
• Using tools and function-calling
• Introduction to LangChain Expression Language (LCEL)
• Hands-on: Use LangChain to query a dataset using chains & agents

• What is ReAct? Reasoning + Acting loop
• Differences between ReAct, CoT, and Reflexion
• Reflection and Reflexion: Iterative self-correction
• Comparing: ReAct vs Reflexion vs ToRA vs LATS
• Hands-on:
  • Implement a ReAct agent in LangChain
  • Implement a Reflexion agent using LangGraph state

• Introduction to LangGraph (State Machines + LangChain)
• Nodes, edges, states, loops, branching logic
• Designing Agentic workflows with memory
• Reflexion loops using LangGraph
• Hands-on:
  • Build a LangGraph agent with self-reflection
  • Multi-hop RAG using LangGraph

• What is CrewAI? How it compares to LangGraph
• Agents, Roles, Tasks, Crew
• Tools and Custom Agents
• Planning multi-step agent flows
• Hands-on:
  •  Build a 3-agent CrewAI system (e.g., Question Generator → Retriever → Answer Synthesizer)

Project: Build an Agentic RAG System using CrewAI and LangGraph

• Project Brief:
  “Create a system that takes a document, extracts multiple questions, and answers them
  using Agentic RAG.”
• Phase 1: Document upload + RAG indexing
• Phase 2: Question generator agent (CrewAI or LangGraph node)
• Phase 3: RAG agent + synthesizer
• Phase 4: Reflexion loop for better answer quality
• Optional: Gradio front-end or Streamlit demo