Pulse

A fitness problem with no object yet

We stalled. Lots of problem, no product. I floated directions until a smart mirror felt right, a coach in the room instead of another app. We talked it through until we had something we could build.

Interviews together, synthesis after

I pulled Crazy 8s and expert interviews into a plan we could design against: feature priority, how dense the mirror UI should be, how you interact, and where it lives in a dorm.

Eight sketches and three experts

Crazy 8s put an early workout-assistant mirror on paper. Yoga, health-tech, and fitness instructors told us what a coach in the room would actually have to do.

Marin falls off after one miss

I wrote Marin as the person we were designing for: a software engineer in SF who loses the gym the week after she skips one session.

AirTap, because sweat is a constraint

Useful, and a little invasive

We concept-tested calendar sync, facial tracking, and posture coaching with five people at CCA. Calendar-aware micro workouts landed. Camera tracking was useful and a little invasive. That tension is still in the prototype on purpose.

From deck to something you can play

The rest timer collided with navigation. The gesture logic was clever enough to exit a workout by accident.

We all ran sessions. Then I looked at the pile.

One gesture, not a switch

People hated switching between point and pinch. They wanted one gesture. Tracking was jumpy enough to exit a workout by accident.

Onboarding that talks too much

Too much text up front, dead rest screens, and exercises that were hard to identify from the mirror.

Honest about fit