I’m Matt.

The part I enjoy most

I like getting into a complicated journey: following what people do, examining the data, and working through the details until the problem starts to make sense. There is a particular satisfaction in the moment when something that looked confusing becomes clear enough to act on.

At Cafe Bazaar, the total number of Inline Apps users did not explain what was happening inside each category. Looking more closely at that behavior helped us focus on video. At Sotoon, time with quality managers helped us understand what they needed to do after a call had been transcribed. In both cases, getting closer to the work changed the product direction.

Working with other people

I value clear expectations and direct feedback. I want to know where my reasoning is weak while there is still time to do something about it. With colleagues, I try to make the decision clear, explain the trade-off, and leave room for someone to change my mind.

I still want to be close enough to the details to understand what I am asking a team to build and why.

Thinking with AI

I use AI a lot. One thing that concerns me is how easily a fluent answer can make an idea sound better examined than it is. I ask for weak assumptions and counterexamples, but asking for criticism does not automatically make the answer reliable. I still have to check it.

My writing and tools explore related questions: what evidence makes an AI feature ready, which actions a system should be allowed to take, and how people can review or correct it.