51.514 Applied Data Privacy for AI Systems

Course description

This course equips students with the knowledge and hands-on skills to design and implement privacy-preserving AI systems in real-world contexts such as healthcare, finance, and smart nation applications. Students learn the core building blocks of privacy-preserving AI, including differential privacy, federated learning, and secure multiparty computation. Through practical labs using open-source tools and frameworks, students will implement and evaluate privacy techniques under realistic constraints such as utility trade-offs, governance needs, and deployment considerations. Students then apply these concepts in industry-style projects based on common challenges faced by public and regulated sectors, including sensitive data sharing, model monitoring, and compliance-driven design. The course also introduces selected industry case studies and practitioner perspectives through guest sessions or partner engagements where available.

Instructor

Victor Keong

Number of credits: 12