At Meta, I brought user research into AI model development across four product areas. I established user-defined measures of helpfulness, trust, and intent fulfillment, contributing to a 30% increase in user satisfaction.
I developed AI evaluation frameworks using interviews and regression analysis, and assessed agreement between human raters and LLM judges. My research on a 0-to-1 AI product informed feature priorities and product direction, contributing to a 2x increase in adoption.
At 23andMe, I co-first-authored a longitudinal study on the psychological safety of communicating genetic risk for depression. I also led participant engagement work for a research cohort of 12M+ individuals, with five personalized campaigns resulting in 2x participation rates.
At BetterUp, I designed and evaluated a measurement-based coaching intervention whose participants showed 2x greater improvement in well-being outcomes than the control group. Across my research work, I use experimental design, behavioral data, and mixed methods to inform product and scientific decisions.

