Talks & Conversations

I write and speak about decision systems, evaluation under uncertainty, and how scientific organizations actually adopt AI — beyond the demo, into the operating layer.

Most of my time goes into the day job. The rest I spend trading notes with peers in data science, product, and research leadership who are working through similar problems.

What I’m thinking about

These are the conversations I don’t get tired of:

• Evaluation when ground truth is delayed, partial, or contested. What counts as evidence, and what’s metric theater dressed up as rigor.
• Decision-support tooling that scientists actually use. Instead of bypass with a side of resentment.
• Applied AI inside scientific organizations. The models are the easy part — the operating layer is where it lives or dies.
• The translation layer. Research rigor on one side, product judgment on the other, and the people in between trying to ship.

If your team is working through any of these and would benefit from an outside perspective — for a talk, a panel, a guest session, or a longer conversation — drop a note.