Built for the hard parts of AI delivery.

The gap between AI research and AI in production is not technical. Teams that can build remarkable things still stall out because they have to wait on approvals that were never clearly assigned, blocked on dependencies no one owns, shipping to a definition of done that nobody agreed on at the start.

I got into program management because I wanted to work on the things that actually matter: the decisions, the alignment, the accountability structures that determine whether a program ships or stalls. AI made that work more complex, higher stakes, and more interesting.

Over the past decade, I have worked forward deployed with enterprise teams at different stages of the AI adoption curve, with some just starting to think about agents, others sitting on a $25M portfolio of programs with unclear ownership and no governance model.

The work is always some version of the same thing: get the structure right, get the right people aligned, then get out of the way of the team.

Engineering, product, legal, and security rarely speak the same language when AI is involved. I translate between them not by simplifying the technical reality, but by building the shared vocabulary and decision making structures that let all four functions move together.

I do not manage by dashboard. Governance frameworks and reporting cadences are tools, not the work itself. The work is being in the room when the hard calls get made and making sure those calls have the right information and the right people behind them.