In practice that means machine learning systems built end to end, from exploratory analysis and feature engineering through production-ready deployment with containerized infrastructure, monitoring, and CI/CD. I recently trained a CatBoost model that predicts used-car prices within ~$1,300 at the median across 243,000 vehicles and packaged it as a production-ready API with FastAPI, Docker, Kubernetes, Prometheus and Grafana monitoring, and automated testing through GitHub Actions. Around that sit data pipelines, geospatial analysis, and client-facing applications, including a live tool that turns a client's demolition site list into a complete neighbor notification package in about a minute.

What I care about most is the edges, where real inputs meet a system that expects clean ones. Sometimes that means input handling, so a typo returns a corrected estimate instead of a rejection. Sometimes it means deciding which kind of error is acceptable and routing the ambiguous cases to a person rather than quietly dropping them. On the demolition tool, missing an occupied home is not comparable to printing an extra door hanger, so the filter errs toward including, and every run ships an assumptions log showing how each address was derived.

That instinct came from the other side of the table. As Director of Strategic Planning for a national retail company, I oversaw marketing investments, store expansion, and promotion optimization, making calls off tools whose limits I could not see and contractor reports I could not easily interrogate. Being handed a number nobody can explain is one lesson. Watching people use a system in ways nobody designed for is the other. I know what data-driven strategy looks like from the decision-maker's seat and not just the analyst's, and I build for the people who inherit the tool rather than the person who wrote it.

Curiosity is the other half of it. I've built and tuned AI models, used optimization tools like Gurobi to identify global minima and maxima for strategic decision-making, and developed production-grade Python applications, earning a 4.0 GPA in my Master of Science in Business Analytics at William & Mary along the way. After completing my Python course there, I stayed on to tutor other students, which sharpened my own fundamentals as much as theirs.

I earned my undergraduate degree from Georgetown University's School of Foreign Service, majoring in Science, Technology, and International Affairs. Before studying Mandarin Chinese formally, I had already begun teaching myself the language and passed early fluency exams. At Georgetown, I deepened that foundation and passed an advanced fluency exam, a reflection of the same persistence and self-directed learning I bring to every technical challenge. Having lived in Jackson Hole, Wyoming, Northern Virginia, and China, I've developed an appreciation for diverse perspectives and an ability to collaborate across cultures.