At work, those questions run at a different scale.
The organization spans government, military, FFRDC, and engineering partners, with work across generative AI, computer vision, predictive analytics, automation, and mission software.
The goal isn’t simply to build prototypes. We’ve developed more than 50 user-driven capabilities and transitioned roughly 10 into enduring program software, while building enterprise AI patterns that connect mission applications, tools, and hundreds of underlying data sources.
A big part of my job is closing the gap between technology and the people who actually need it. That includes embedding technologists with operational organizations for months at a time, shaping major acquisitions, and helping senior government leaders understand which models, architectures, and investments are ready to matter.
At that scale, putting a chatbot in front of the data is the easy part. The actual system has to find the right information, use the right tools, understand context, respect security boundaries, and produce answers people can trust.
The projects on this site give me firsthand evidence to carry back into that work: where agents lose context, where generative interfaces misstate data, which controls belong in code, and what useful developer tooling feels like. So when I sit with engineers, technology companies, or government leaders, the discussion starts from systems I use, test, and break — not a slide deck.
