Recent Projects: How we estimate the risk from prompt injection attacks on AI systems; Protecting users with differentially private synthetic training data.
Background
I lead the Security and Privacy Research team at Google DeepMind, working on privacy and security for ML systems. Before joining Google I was an associate professor in the Department of Computer Science at the Johns Hopkins University, where I headed the Hopkins InterNetworking Research (HiNRG) Group. I worked on computer networks with an emphasis on low-power and sensor networks.