Full course coming soon

Hardware Architecture for AI System Design

7 chapters · 33 lessons

1. 1. Semiconductor Fundamentals for AI Practitioners

   4 lessons

   Understand how chip design, fabrication, and packaging create the performance envelope for AI workloads.

2. 2. Memory Hierarchies and Data Movement Bottlenecks

   5 lessons

   Analyze how memory capacity, bandwidth, and latency constrain model size, batch size, and throughput.

3. 3. Compute Architectures: CPUs, GPUs, and AI Accelerators

   5 lessons

   Compare processor types and identify which workloads benefit from specialized acceleration.

4. 4. Interconnects, Bandwidth, and System Integration

   4 lessons

   Evaluate how data transfer between components affects end-to-end system performance.

5. 5. Power, Thermal, and Efficiency Constraints

   5 lessons

   Design within power budgets and thermal limits for data center, edge, and mobile AI deployments.

6. 6. Hardware Selection and Procurement Strategy

   5 lessons

   Make informed decisions about chip selection, vendor evaluation, and supply chain risk.

7. 7. Optimization Techniques Across the Hardware Stack

   5 lessons

   Apply software and configuration strategies to extract maximum performance from existing hardware.