At micro1, I evaluate the quality and suitability of materials used in AI training and review whether scientific claims match their cited evidence. In my independent research, I pre-registered a pan-cancer analysis plan with an estimand and negative controls.
As analytics lead for the White House COVID-19 H-CORE test-kit mission at Booz Allen Hamilton, I led analysts producing allocation numbers for 1 billion COVID-19 test kits. I chose the CDC/ATSDR Social Vulnerability Index over a custom model so the allocation rule would be explainable to HHS and usable the same month.
At Geisinger Health System, I partnered with clinicians to study heart-failure genetics using exomes and longitudinal EHR phenotypes. My work draws on mixed-effects models, survival analysis, and analysis planning, with R and SAS.
