AI-powered training tool for Team USA

Feels Like partnered with Google Cloud ahead of the Olympic Winter Games, helping design and build a first-of-its-kind AI training tool for the US Ski & Snowboard freestyle and snowboard teams.

Turning feel into data

Skiing and snowboarding run on feel: flow, muscle memory and instinct. But what if feel could become data, giving coaches and athletes a more precise, comparable way to train?

A model at home in the slopes

Google Cloud and DeepMind built custom models trained to understand athlete motion and the specific demands of snow sports. The result: a tool that turns any slope into a biomechanics lab, using footage captured on conventional smartphones. The interface designed for two states: In device on the slopes, mid-session, gloves on; and a larger screen at basecamp for the debrief.

What does a cork 1440 look like?

We designed new ways to see a trick. The Cork Ribbon maps the spin axis through the air. The Muybridge strip lays the trick out in one dimension along a timeline. The full arc of a run readable at a glance. The skeletal overlay captures the body's mechanics frame by frame: the compression into the takeoff, the tuck, the extension on the landing.

Telling the story

We also designed and built the Google Cloud case study website that brought the project to the world, rich with animation, video, and athlete profiles.

CREDITS

Feels Like


Design, Engineering