CSC 375/575 - Course

About the Course

CSC 375/575 Generative AI is an advanced course that takes students from foundational LLM concepts to cutting-edge large-scale AI system implementation. Starting with Raschka's hands-on approach to building language models from scratch, students master tokenization, attention mechanisms, and transformer architectures. The course then advances to large-scale training systems using Xiao & Zhu's comprehensive framework, covering distributed training, efficient attention variants, and memory optimization techniques. Advanced topics include systematic prompt design, chain-of-thought reasoning, retrieval-augmented generation (RAG), reinforcement learning from human feedback (RLHF), constitutional AI, and inference-time scaling. Students complete hands-on projects implementing core LLM components and develop expertise in deploying production-ready generative AI systems.

Course Schedule

Course Materials

📄 Course Syllabus

Available Lectures

Lecture	Topic	Materials
1	Introduction to Generative AI	Slides Handout
—	PyTorch	Google Colab (Part 1) Google Colab (Part 2)
2	LLM Foundations & Pre-training	Slides Handout Notebook Google Colab
3	Tokenization & Data Processing	Slides Handout Notebook Google Colab
4	Attention Mechanisms & Transformers	Slides Handout Notebook
5	Building GPT Architecture: Implementing Core Model Components	Slides Handout Notebook Google Colab
6	Model Training Pipeline: Pre-training Large Language Models from Scratch	Slides Handout Notebook Google Colab
7	Fine-tuning for Text Classification	Slides Notebook Google Colab
8	Instruction Fine-tuning: Aligning Models with Human Instructions	Slides Handout Notebook Google Colab
9	Parameter-Efficient Fine-tuning with LoRA	Notebook Google Colab
10	Prompting Techniques and Chain of Thought	Notebook (Part 1) Notebook (Part 2) Notebook (Part 3) Google Colab (Part 1) Google Colab (Part 2) Google Colab (Part 3)
11	Alignment: SFT, RLHF, and DPO	Notebook (Part 1) Notebook (Part 2) Google Colab (Part 1) Google Colab (Part 2)

Assignments

Assignment	Topic	Materials
1	Building Meta's LLaMA Tokenizer	Download Package Instructions Submit
2	Building GPT-2 from Scratch - Enhanced with Advanced Concepts	Download Package Instructions Submit
3	Classification and Instruction Fine-Tuning	Download Package Instructions Submit
4	Prompt Engineering Experiments	Instructions (Submit with Final Project)
Final	Semiconductor Simulation Code Generation	Download Dataset Instructions

Course Structure - 5 Progressive Learning Phases

Phase 1: Foundations (Weeks 1-4) - Raschka Ch.1-6

Build LLMs from scratch: tokenization, attention mechanisms, transformer architecture, pre-training, and supervised fine-tuning

Phase 2: Core Implementation (Week 5) - Raschka Ch.7

Master instruction fine-tuning techniques to align models with human instructions and preferences

Phase 3: Large-Scale Training Systems (Weeks 6-10) - Xiao&Zhu Ch.2

Advanced training infrastructure: scaling laws, distributed training, efficient attention variants, and memory optimization for production systems

Phase 4: Prompting & Tool Integration (Weeks 11-12) - Xiao&Zhu Ch.3

Systematic prompt design, chain-of-thought reasoning, RAG systems, and automatic prompt optimization techniques

Phase 5: Alignment & Inference Optimization (Weeks 13-14) - Xiao&Zhu Ch.4-5

RLHF implementation, constitutional AI, human preference learning, efficient inference, and inference-time scaling

Currently Available: Lectures 1-7 covering foundational concepts through GPT architecture, pre-training, and fine-tuning. Additional lectures will be released progressively throughout the semester.