I'm a data scientist whose work touches complex problems across the applied sciences: algorithms, artificial intelligence and machine learning, and the mathematical modeling underneath them. Day to day, that means reading and reasoning about the source code of complex systems, building statistical and machine-learning models from large and often messy datasets, and turning both into analysis rigorous enough to stand entirely on its own.

As Vice President, Data Science at Analysis Group, I lead technical teams through some of the most demanding quantitative work in litigation: forensic source code review of complex software systems, large-scale econometric modeling, and NLP-driven algorithm design. I work as an expert witness, and I hold my own analysis to a simple standard: methodology that is reliable, testable, and rigorously applied to the facts, not just asserted. That's just good science.

That technical core is quantitative to its bones: statistics, machine learning, optimization, and the algorithms that make them tractable at real-world scale. It has taken me into algorithmic collusion, into evolutionary algorithms that design neural network architectures on their own, and into the forensic work of reconstructing exactly what a piece of software did using nothing but the code and data it left behind. Lately I've been drawn to a broader question: what agentic AI systems can actually do inside serious quantitative work, not just automate around it. One thread I keep pulling on is large language models as tools for statistical modeling itself, feature extraction in econometric analysis in particular, which feels like a genuine shift in how applied economists will build models next.

I'm also an adjunct professor at Boston College, where I teach graduate courses in machine learning, econometrics, programming, and data analysis. The math underneath these methods is too often taught as something dry and abstract: equations on a board, disconnected from what they actually do to data, though for the record, I love a good board full of equations (see below). This lab is my attempt to build the machinery that makes that same math tangible: tools you manipulate and watch respond, so the ideas stop being abstract and start being something you can see.