Konnor Herbst

I’m broadly interested in connecting modern machine learning methods with classical statistics: using flexible methods while retaining valid inference and defensible scientific conclusions. My methodological interests include causal inference, treatment policies, treatment effect heterogeneity, modern nonparametric methods, experimental design, and methods for multiple or prioritized outcomes. My applied interests include cardiovascular and aging-related diseases, drug repurposing, and program evaluation in education, clinical care, and public health practice settings.