Research Portfolio
From early psychophysiological driver studies to large‑scale ML‑assisted survey analytics and org‑level research operations, this portfolio highlights a quantitative + mixed‑methods UX research practice focused on scalable insight pipelines, methodological rigor, and adoption-aware process design. Each project is framed with the STAR model to surface decision context, execution leverage, and measurable impact.
Led a longitudinal study tracking ML model performance pre/post-launch; built a 6-step GenAI classifier pipeline (Self-consistency, LLM-as-a-judge) that reduced analysis time ~73% and was adopted by the wider research team.
Real‑world driver stress elicitation: replicated 90% stress induction & 89% event correlation via mixed-methods instrumentation.
Framework + longitudinal HW usability evaluation driving remote redesign & metric expansion across millions of devices.
Progressive rollout of standardized repository & reporting boosted visibility and self‑serve insights org‑wide.