AWS SageMaker

Interaction design for AWS's machine-learning platform

A machine-learning platform serves six kinds of people. Data scientists, ML engineers, analysts, developers — each works the same models from a different angle. How they'd actually collaborate inside the interface was an open question.

SageMaker needed a collaboration model nobody had drawn yet. Working from the team's own personas, I prototyped the options, low- and high-fidelity interactive flows, synchronous and asynchronous, that stakeholders could click through and feel instead of debating in the abstract.

I never found out which patterns shipped. That wasn't the job. It was to make the options concrete enough to choose between. UX for an AI platform, before AI was a big deal.