OATML

Xander is a PhD student in the OATML group, co-supervised by Yarin Gal and Sadie Creese. Xander was on the founding team of the UK AI Safety Institute, where he founded and led the adversarial ML team (Safeguards Analysis), performing adversarial testing on a range of frontier AI systems on behalf of the UK government. Xander’s main research interests are attacking and understanding safeguards of frontier AI systems. Previously, Xander studied Computer Science at Harvard, where he founded the Harvard AI Safety Team.