Skip to content

course_outline.md

Latest commit

 

History

History
196 lines (163 loc) · 7.27 KB

course_outline.md

File metadata and controls

196 lines (163 loc) · 7.27 KB

Hands-On Generative Artificial Intelligence

Last Updated: February 23, 2026
Duration: 3 Days

Course Overview

This intensive three-day training course is designed for developers, data engineers/analysts, and tech product owners. It combines theory with extensive hands-on practice to teach participants how to build production-ready generative AI applications using state-of-the-art models and techniques.

Participants will gain practical experience with modern LLMs, fine-tuning strategies (including LoRA/PEFT), retrieval-augmented generation (RAG), AI agents, and production optimization techniques. The course includes 32 hands-on labs covering the complete lifecycle from model development to deployment.

Prerequisites

• Python Programming: Solid understanding of Python, including data structures, control flow, functions, and libraries like NumPy and Pandas
• Machine Learning Fundamentals: Familiarity with supervised/unsupervised learning, model evaluation, and scikit-learn
• Deep Learning Basics: Understanding of neural networks recommended but not required
• API Experience: Helpful to have worked with REST APIs (not required)

Course Outline

Day 1: Foundations and Generative AI Basics

Module 01: Foundations of AI and Machine Learning

• Machine Learning vs rule-based programming
• Supervised and unsupervised learning with real-world applications
• ML model development lifecycle:

  • Data preprocessing and feature engineering
  • Overfitting and model evaluation metrics • Hands-on Lab: Training and evaluating a classifier

Module 02: Deep Learning Primer

• Fundamental concepts of deep learning
• Neural network architectures and common patterns
• Optimizers, gradient descent, and backpropagation
• Deep learning frameworks: TensorFlow and PyTorch
• Hands-on Labs:

  • Neural network basics
  • Image classification using PyTorch

Module 03: Overview of Generative AI

• Introduction to Generative AI and its applications
• Evolution of generative models:

  • Autoencoders and latent space representation
  • Variational Autoencoders (VAEs) and probabilistic sampling
  • Diffusion Models for high-fidelity image generation • Hands-on Labs:
  • Autoencoders with MNIST dataset
  • Introduction to Diffusion Models

Day 2: NLP, Large Language Models, and Intelligent Agents

Module 04: NLP - Understanding Language as Data

• Introduction to NLP techniques and applications
• Tokenization and text preprocessing
• Vectorization: Bag-of-Words and limitations
• Embeddings: Word2vec and semantic representations
• Hands-on Lab: Finding similar documents using word2vec

Module 05: Large Language Models (LLMs)

• Evolution of LLMs: From BERT/GPT-2 to modern models
• Current popular models:

  • Closed-source: GPT-4, Claude, Gemini
  • Open-source: Llama 3, Mistral, Phi-3 • Model architectures and selection criteria
    • Hugging Face ecosystem and model hub
    • Multimodal Models (Vision-Language Models) • Building interactive UIs with Gradio
    • Hands-on Labs:
  • Using OpenAI and Ollama APIs
  • Exploring Hugging Face model hub
  • Comparing BERT and GPT models
  • Creating interactive UIs with Gradio
  • Exploring Multimodal Models

Module 06: Prompting Techniques and Agentic AI

• Advanced prompting strategies:

  • Few-shot learning and in-context examples
  • Chain-of-Thought (CoT) prompting
  • Structured prompting frameworks • Function calling and tool use
    • Building AI agents:
  • ReAct (Reasoning + Acting) pattern
  • LangChain agent frameworks
  • Autonomous workflows • Hands-on Labs:
  • Systematic prompt optimization
  • Function calling with OpenAI
  • Function calling with LangChain
  • Building ReAct agents

Day 3: RAG, Fine-Tuning, and Production Optimization

Module 07: Retrieval-Augmented Generation (RAG)

• RAG architecture and use cases
• Vector databases and semantic search:

  • Embeddings for document retrieval
  • Chroma vector database • Building RAG pipelines:
  • Document chunking and preprocessing
  • Embedding generation and storage
  • Retrieval strategies and context injection • RAG evaluation and observability:
  • Evaluating RAG with LLM-as-a-judge
  • MLflow for LLM tracking
  • Quality metrics and debugging • Hands-on Labs:
  • RAG with LangChain and Chroma
  • RAG with LlamaIndex
  • RAG evaluation with LLM-as-a-judge and MLflow

Module 08: Fine-Tuning Large Language Models

• Transfer learning strategies for LLMs
• Parameter-Efficient Fine-Tuning (PEFT):

  • LoRA (Low-Rank Adaptation)
  • QLoRA (Quantized LoRA)
  • Full fine-tuning vs PEFT comparison • Catastrophic forgetting and mitigation strategies
    • Sampling techniques:
  • Temperature, top-p, top-k
  • Beam search • Hands-on Labs:
  • Transfer learning concepts
  • Sentiment analysis with DistilBERT
  • Fine-tuning with OpenAI API
  • Summarization fine-tuning
  • Sampling techniques exploration
  • LoRA fine-tuning for healthcare applications

Module 09: Model Optimization and Deployment

• Production challenges: memory, cost, latency
• Optimization techniques:

  • Knowledge distillation (teacher-student training)
  • Model pruning (structured and unstructured)
  • Quantization (FP16, INT8, INT4, GPTQ, AWQ) • Benchmarking and performance evaluation
    • Deployment strategies:
  • Cloud deployment (GPU/CPU)
  • Edge deployment
  • Production optimization trade-offs • Hands-on Labs:
  • Introduction to optimization
  • Knowledge distillation
  • Model pruning
  • Quantization for deployment
  • Benchmarking optimized models

Module 10: Capstone Project

• Building a production-ready dialogue system with RAG
• Integrating multiple techniques learned throughout the course
• Best practices for deploying GenAI applications

Timing Guide

Day 1 • Module 01: Foundations of AI and Machine Learning
• Module 02: Deep Learning Primer
• Module 03: Overview of Generative AI

Day 2 • Module 04: NLP - Understanding Language as Data
• Module 05: Large Language Models (LLMs)
• Module 06: Prompting Techniques and Agentic AI

Day 3 • Module 07: Retrieval-Augmented Generation (RAG)
• Module 08: Fine-Tuning Large Language Models
• Module 09: Model Optimization and Deployment
• Module 10: Capstone Project

What You'll Learn

By the end of this course, you will be able to:

✅ Build applications with modern LLMs (GPT-4, Claude, Llama, Mistral)
✅ Fine-tune models efficiently using LoRA and PEFT techniques
✅ Implement RAG systems with vector databases (Chroma)
✅ Create AI agents with function calling and ReAct patterns
✅ Optimize models for production (distillation, pruning, quantization)
✅ Deploy GenAI applications with proper evaluation and monitoring
✅ Use industry-standard tools (Hugging Face, LangChain, LlamaIndex, MLflow)

Course Features

• 32 Hands-on Labs: Practical notebooks covering every major topic
• Production-Focused: Learn deployment and optimization techniques
• Modern Tools: Work with 2026 industry-standard frameworks
• Complete Lifecycle: From model selection to production deployment
• Real-World Projects: Build a complete RAG-based dialogue system