May Dixit

May is a Research Scientist on the Frontier AI Risk Management team, building evaluations and risk assessments for frontier models. Before transitioning to AI safety research, she spent nearly a decade as an engineer, including six years as an ML Engineer at Google working on classifier training and ML fairness. Most recently, as a Research Fellow at Cambridge ERA, she studied the effect of agentic post-training on model propensities toward misalignment. May holds an MS in Applied Mathematics from Stanford University.