This course covers the practical implementation of machine learning systems at scale. Students will learn how to build ML systems that are efficient, robust, and deployable across diverse hardware and software environments. Topics include model optimization, memory and compute management, GPU acceleration, distributed training, quantization, inference serving, and evaluation of system trade-offs. Assignments emphasize hands-on coding in modern frameworks, benchmarking, and profiling. The final project requires students to implement and optimize an ML system (e.g., training pipeline, serving infrastructure, or deployment on edge devices) and present empirical results. This project-oriented approach ensures students gain both theoretical understanding and real-world engineering experience.

This course is a study of the hardware and software issues in parallel computing. Topics include an introduction to the basic concepts, parallel architectures and network topologies, parallel algorithms, parallel metrics, parallel languages, granularity, applications, parallel programming design and debugging. Students will become familiar with various types of parallel architectures and programming environments.