Data Scientist — Environmental & Natural Resource Systems
Building data tools for federal natural resource work at USDA Forest Service. Background in healthcare AI — FHIR, Epic, clinical NLP. Most of what I write down is about measuring whether a result survives being checked a second way.
Environmental data science supporting federal natural resource initiatives.
Early sepsis prediction from ICU time series, built to measure its own failures rather than assert their absence. Moving the model from medical to surgical ICU costs almost no AUROC and 91% of its clinical value — discrimination and the operating point fail independently, and one external AUROC column hides both.
Built LLM-powered tools for clinical research and healthcare delivery.
Progressed from administrative to project management roles, building healthcare systems expertise across UNC research programs.
A sepsis model crossed from medical ICU to surgical ICU with its AUROC almost intact and lost 91% of its clinical value. Discrimination and the operating point fail independently, and the one number everyone reports is insensitive to the failure that costs the most.
I spent a week measuring a biomedical knowledge graph system. Five separate results looked solid at a single configuration and dissolved under a second one. Here is each mechanism, and the one that nearly shipped as a feature.