I start from what could go wrong, working edge-case-first helps me think more clearly, and the ideal experience usually follows from there.

I prototype in code because static designs can't show you everything; state transitions, edge cases, interaction patterns all feel different when they're actually running.

I ask a lot of questions, including the ones that feel obvious. Most of the time, those are the ones worth asking.

Before anything ships, I run it through a few rounds of dogfooding and usability testings, without a miss.

I care deeply about shipping quickly, validating with real users, and learning through iteration rather than perfecting in isolation.