Santiago Zanella-Beguelin

I'm a Senior Researcher in Computer Science at Microsoft Research focusing on privacy and security in machine learning systems. My research interests span differential privacy, machine learning privacy, programming languages, and formal verification. I work on developing theoretical frameworks and practical tools to analyze and enhance the privacy guarantees of ML systems. My recent work has focused on analyzing privacy risks in large language models and developing techniques for measuring and bounding information leakage in machine learning models. I've collaborated on organizing the SaTML LLM Capture-the-Flag competition to study security risks in LLM systems. I'm particularly interested in developing closed-form bounds and theoretical frameworks for evaluating privacy guarantees in differentially private machine learning.

I have a strong focus on bridging theoretical foundations with practical applications. My research aims to develop rigorous methods for quantifying and protecting privacy in modern machine learning systems while maintaining their utility. I collaborate extensively with researchers across academia and industry to advance the field of privacy-preserving machine learning.