Generative AI

Learn Generative AI, Large Language Models (LLMs), AI automation, and intelligent AI systems using real-world datasets, modern AI frameworks, and hands-on practical implementation.

Generative AI image
AI

What is Generative AI?

Generative AI is the process of creating intelligent systems that can generate text, images, code, audio, and digital content using advanced AI models and large-scale datasets.

Generative AI

Module 1 : Introduction & Foundations
  • Introduction
  • What is Generative AI?
  • What is Agentic AI?
  • Generative AI vs Agentic AI

Module 2 : Fundamentals of Generative AI & LLMs
  • Basics of Generative AI
  • What are LLMs?
  • Key Concepts
  • Tokenization
  • Embeddings
  • Vocabulary
  • Attention
  • Transformer Architecture
  • Self-Attention Mechanism
  • Multi-head Attention
  • Cross Attention
  • Encoder–Decoder

Module 3 : Working with LLMs
  • Open Source vs Closed Source LLMs
  • Hugging Face Ecosystem
  • Model Loading
  • Model Parameters
  • Model Weight Formats (.pth, safetensors, onnx)
  • Model Size Calculation and License
  • Multimodal LLMs: Text, Audio (ASR, TTS), Image, Video

Module 4 : Prompt Engineering & Control
  • Prompts and Context
    • Max Sequence Length vs Max Output Tokens
    • Task-specific Prompts
    • Sampling Parameters (Temperature, Top-k, Top-p, etc.)
  • Prompting Techniques
    • Zero-Shot Learning
    • Chain of Thought (CoT)
    • ReAct
  • Guardrails
    • Using Chat Completion APIs
    • OpenAI (ChatGPT)
    • Google Gemini
    • Anthropic Claude

Module 5 : Retrieval-Augmented Generation (RAG)
  • What is RAG?
  • LLM Hallucination: Causes and Mitigation
  • When to Use RAG
  • Components of RAG
    • Embeddings
    • Vector Databases: Chroma, Pinecone, FAISS, Milvus
    • Chunking Strategies
    • Conversational RAG
    • Embedding Spaces: Semantic Similarity & Cosine Distance
    • Answer Grading / Response Evaluation (BLEU, ROUGE, GPT-based)

Module 6 : Advanced RAG Techniques
  • Corrective RAG (CRAG)
  • Self-RAG with Reflection
  • Graph RAG
  • Hybrid RAG(Semantic + Keyword)

Module 7 : Graph-based Knowledge and Retrieval
  • Graph Fundamentals – Nodes, Edges
  • Ontology Design
  • GraphDBs:
    • Adavantages of GraphDBs
    • Neo4j:Community vs Enterprise vs Cloud

    • Module 8 : LangChain Framework
    • Introduction to LangChain
    • Chains, Prompts, and Templates
    • Memory Systems and Conversation Flow
    • Memory Types: Short-term, Long-term, Episodic
    • Persistence Strategies
    • Basic Document Loaders and Text Splitters

    Module 9 : Agentic AI Principles
    • What is Agentic AI?
    • AI Agents vs Agentic AI
    • Agentic AI Frameworks Overview
      • No-code vs Code-first
      • N8n vs LangGraph vs Airflow
      • Design Principles:
      • Goal,Planner,Orchestrator
      • Copilot vs Autopilot
      • Agentic AI Frameworks
        • CrewAI, LangGraph, AutoGen

    Module 10 : Production & Final Project
    • Agentic AI using LangGraph
    • LangChain vs LangGraph
    • LangGraph Components
    • Workflow Types
      • Sequential
      • Parallel
      • Iterative
      • Conditional
    • Memory & State Management
      • Persistence Strategies
      • Time Travel in LangGraph
    • Observability with LangSmith
      • What is Observability?

    Module 11 : Model Context Protocol (MCP)
    • MCP Fundamentals and Architecture
    • MCP Server and Client Implementation
    • Tool Integration through MCP
    • MCP vs Traditional Integration Patterns

    Module 12 : Agent-to-Agent Communication
    • Agent-to-Agent Communication Fundamentals
    • Orchestration Patterns:
      • Manager-Worker
      • Peer-to-Peer

    Module 13 : Fine-Tuning and Quantization
    • When to Use Fine-tuning vs Prompt Engineering
    • Parameter-Efficient Fine-Tuning (PEFT)
    • LoRA
    • QLoRA
    • Quantization Techniques
      • Intro to Quantization
      • Asymmetric vs Symmetric
      • Post-training Quantization vs Quantization-Aware

    Module 14 : Model Serving & Deployment + Projects
    • Model Serving Frameworks:
      • vLLM
      • TensorRT-LLM

    • Final Projects (Pick One)
    • Domain-Aware LLM Chatbox using open-Source
    • Customer Support Assistant Powered by Retrieval-Augmented Generation
    • Agentic AI Advisor for Healthcare Guidance and Decision support