Course Overview

This 5000-level postgraduate course introduces state-of-the-art computer architectures, networks, and distributed systems for AI computing. It connects hardware and systems evolution with large language model infrastructure, then brings the ideas together in an end-to-end course project.

Prerequisites

• Operating Systems, Design and Analysis of Computer Architectures, or Computer and Communication Networks
• Basic C++ programming knowledge
• Basic machine learning knowledge
• A passion for understanding computer architectures, networks, and LLM systems in depth

Major Learning Goals

• Understand widely used and state-of-the-art techniques in AI infrastructure.
• Design and build an end-to-end infrastructure stack for large language models.
• Develop practical skills in multithreaded programming and AI algorithms that are valuable for research and industry careers.
• Build intuition for system design choices, trade-offs, and the many interconnected factors involved in becoming a strong system architect.

Lecture Schedule

Academic Integrity

All exercises submitted for grading must represent the results of each student's own independent efforts. Students may discuss methodologies with classmates, but must do the work independently.

• Do not copy work from another student or person.
• Do not look at another student's actual solution.
• Do not share actual code or solution schemes with other students or people.
• Do not post your work online.
• Discussion is encouraged, but every student must complete the work independently.

Refer to the University's Honor Code and the University's penalties for cheating.