Two practices, one way of paying attention.

I am a registered patent agent. For five years my days have been claim charts, office actions, and the particular vocabulary that turns a model into an invention the Office can examine. Most of that work is in artificial intelligence and machine learning: architectures, training methods, data pipelines, deployment.

I went back for an MSc in computer science because arguing about systems you have not built is a bad habit. It finishes this year. The result is less glamorous and more useful than it sounds: I can read an inventor's notebook, a training script, and an examiner's §101 rejection in the same afternoon and tell you where the three disagree.

The photographs are not a sideline so much as the other half. Same instinct: look for a long time, then decide what to leave out of the frame. They are kept here at full resolution and without commentary, which is the only way I can stand to show them.