For AI Labs, Evals, and Agent Systems Teams

I am not trying to position myself as a pure research scientist. The more honest claim is that I am very interested in the research-adjacent layer where a team needs product judgment, workflow design, operator thinking, and enough technical fluency to help model capability become a system people can inspect and trust.

That is part of why AI labs, evals work, and agent systems are so interesting to me. Once the question is no longer just whether a model can do something impressive once, the harder work becomes: what should be measured, what should stay deterministic, how should the agent be constrained, what should the operator see, and how do you know whether the system is actually improving.

My background is in product leadership, company building, growth, and technology strategy. I previously led product at Fliff and Mojo and worked in technology strategy at Cerberus Capital. I like bringing that product and operator judgment into AI-heavy environments where the problem is still a little underdefined.

I also like being close enough to the implementation to test ideas directly. That can mean using Claude Code and Codex to build a small tool, mock an operator flow, tighten a spec, create a scoring path, or ship a front-end artifact that helps a team learn faster.

So if you are an AI lab, evals team, or agent systems group that needs someone closer to product builder, workflow operator, or systems-minded generalist than to generic PM process, this is the neighborhood where I think I fit best.