Module 1 – Introduction to Generative AI (8 hours)
- What is Generative AI? Evolution & history
- Difference between Generative AI, Machine Learning, and Deep Learning
- 2025 Trends: Multimodal AI, Agent-based AI, Edge GenAI, Open-source LLMs (LLaMA 3, Mistral)
- Common applications: Text generation, image generation, code generation, video synthesis
- Generative AI workflow
Hands-on:
- • Use ChatGPT and Google Gemini for text-based tasks
- • Prompt engineering basics
Module 2 – Foundations of Large Language Models (10 hours)
- Neural networks recap (transformers, attention mechanism)
- Popular LLM architectures: GPT, LLaMA, Mistral, Claude
- Tokenization & embeddings
- Prompt engineering techniques: zero-shot, few-shot, chain-of-thought
- Latest trend: Function calling & tool-using LLMs
Hands-on:
- • Compare outputs from OpenAI API and open-source LLM (Hugging Face)
- • Design effective prompts for different use cases
Module 3 – Generative AI for Text, Code, and Images (12 hours)
- Text generation: Summarization, Q&A, creative writing
- Code generation with LLMs (GitHub Copilot, Code Llama)
- Image generation: DALL·E, Stable Diffusion, Midjourney
- Latest trend: Multimodal AI (OpenAI’s GPT-4o, Gemini 2.0)
Hands-on:
- • Build a text summarizer using OpenAI API
- • Generate images from text with Stable Diffusion WebUI
Module 4 – Fine-tuning & Customization (10 hours)
- Fine-tuning vs. prompt-tuning vs. adapters (LoRA, QLoRA)
- Dataset preparation for fine-tuning
- Training with Hugging Face Transformers
- Latest trend: Low-resource fine-tuning for edge devices
Hands-on:
- • Fine-tune a small LLM for a domain-specific chatbot
- • Evaluate model performance
Module 5 – Generative AI Deployment & Ethics (10 hours)
- Deploying LLMs on cloud (Hugging Face Spaces, AWS, Azure, Google Cloud)
- Integrating GenAI into apps with LangChain / LlamaIndex
- Integrating GenAI into apps with LangChain / LlamaIndex
- Latest trend: AI watermarking & content authenticity verification
Hands-on:
- • Deploy a chatbot on Streamlit
- • Implement basic AI output filtering
Module 6 – Capstone Project (22 hours)
Project Theme:
- "Generative AI for [Your Domain]" – Learners choose domain (education, healthcare, marketing, entertainment) and:
- 1. Define a problem & dataset
- 2. Select a model (OpenAI, open-source LLM, diffusion model)
- 3. Apply prompt engineering & fine-tuning if needed
- 4. Integrate into an app (Streamlit, Flask, or web UI)
- 5. Deploy online for demo
Examples:
- • AI-powered study assistant
- • Marketing content generator
- • AI recipe generator with images
- • Customer support chatbot with retrieval-augmented generation (RAG)
Final Deliverables:
- • Working application
- • Code + documentation
- • Presentation & demo video
Assessment & Certification
- • Weekly Quizzes: 20%
- • Assignments: 30%
- • Capstone Project: 50%
- • Minimum 60% score for certification
Reviews
There are no reviews yet.