I like difficult ideas best when they become useful systems.

I'm a machine learning engineer with a Ph.D. in computational statistics, but the path connecting those two parts of my background was unusually direct. I worked full-time throughout my bachelor's degrees, master's degree, and Ph.D. I had practical reasons for doing that, including supporting myself and my family, but it also shaped the way I learned.

My education and career continually informed one another. Problems I encountered at work gave me concrete reasons to study statistics, numerical methods, distributed computing, and software engineering more deeply. In the other direction, ideas from coursework and research found their way into production models, data pipelines, optimization systems, and engineering decisions.

Much of my professional work has involved building machine learning products from beginning to end: raw and inconsistent data, feature engineering, statistical modeling, large-scale processing, evaluation, deployment, monitoring, and production operations. More recently, I have extended that background into LLM applications, explainability workflows, and AI agents.

I'm also a father of two boys. Away from work, I spend a lot of time gaming, exploring emulation, and working on my homelab. A server rack in my basement runs a Kubernetes cluster and services used throughout the house. Thin clients connect back to that infrastructure instead of every room needing a separate powerful machine.

One open-source project that brings several of those interests together is Games on Whales. It runs games and virtual desktops in containers on a shared Linux host and streams them to low-latency clients. I use it at home and have contributed commits upstream. It is technically ambitious, practical, open source, and a little playful.