Module 1: Foundations of Generative AI

• Evolution of Natural Language Processing

• Introduction to Neural Language Models

• Generative vs. Discriminative Models

• Types of Generative AI Models

• Introduction to Python Libraries for GenAI (Hugging Face Transformers, OpenAI, etc.)

• Ethical Considerations in Generative AI

Module 2: Transformers and Attention Mechanism

• Attention is All You Need: The Transformer Architecture

• Self-Attention and Multi-Head Attention

• Encoder-Decoder Architecture

• Positional Encoding

• Pre-training and Fine-tuning Paradigm

• Modern Transformer Architectures (GPT, BERT, T5, etc.)

Module 3: Prompt Engineering Techniques

• Introduction to Prompt Engineering

• Zero-shot, One-shot, and Few-shot Learning

• Chain-of-Thought Prompting

• Role Prompting and System Messages

• Temperature and Sampling Strategies

• Structured Output Generation

• Handling Biases and Limitations

Module 4: Token and Cost Optimization

• Understanding Tokenization

• Token Counting and Estimation

• Context Window Management

• Input Chunking Strategies

• Cost Analysis for Different LLM Providers

• Strategies for Reducing API Costs

Module 5: API Integration with LLMs

• Overview of LLM API Providers

• Authentication and API Keys Management

• Making API Calls with Python

• Handling API Responses and Errors

• Streaming Responses

• Rate Limiting and Concurrent Requests

• Project: Creating a Chatbot with OpenAI API

Module 6: Retrieval Augmented Generation (RAG)

• Introduction to RAG Architecture

• Document Processing and Chunking

• Vector Databases and Similarity Search

• Embedding Models for Text Representation

• Query Processing and Reformulation

• Context Augmentation Techniques

• Hybrid Search Approaches

• Project: Building a Question-Answering System with RAG

Module 7: LangChain Framework

• Introduction to LangChain

• Chains and Sequential Processing

• Prompt Templates and Output Parsers

• Memory Types for Conversation Management

• Tools and Agents in LangChain

• Document Loaders and Text Splitters

• Integrating External APIs and Tools

• Project: Building a Multi-Tool Assistant with LangChain

Module 8: PEFT: Fine-tuning & Quantization of LLMs

• Introduction to Parameter-Efficient Fine-Tuning

• LoRA, QLoRA, and Adapter-based Methods

• Dataset Preparation for Fine-tuning

• Model Quantization Techniques

• Deployment Considerations for Fine-tuned Models

• Project: Fine-tuning an LLM for a Specialized Domain

Module 9: Evaluation of LLM Models

• Evaluation Metrics for Generative AI

• Human Evaluation vs. Automatic Metrics

• Benchmark Datasets and Leaderboards

• Evaluating Hallucinations and Factuality

• Building Evaluation Pipelines

• Project: Creating an Evaluation Framework for LLM Outputs

Module 10: Agentic Workflow

• Introduction to AI Agents

• Planning and Reasoning in Agents

• Tool and API Usage by Agents

• ReAct Framework and Chain-of-Thought Reasoning

• Multi-Agent Systems and Collaboration

• Memory and State Management

• Feedback Loops and Self-Improvement

• Project: Building a Task-Oriented Agent

Module 11: Capstone Project - End-to-End Generative AI Application

• Problem Definition and Requirements Analysis

• Architecture Design and Component Selection

• Implementation of Core Functionalities

• User Interface Development (using streamlit or gradio)

• Performance Optimization

• Testing and Evaluation

• Deployment and Documentation

Projects Covered:

• Creating a Chatbot with OpenAI API

• Building a Question-Answering System with RAG

• Building a Multi-Tool Assistant with LangChain

• Fine-tuning an LLM for a Specialized Domain

• Creating an Evaluation Framework for LLM Outputs

• Building a Task-Oriented Agent

• End-to-End Generative AI Application (Capstone)