Yep, exactly. 70% working with colleagues to design ML algorithms (which is mostly thinking of clever ways to extract features, to be honest), and 30% implementing said algorithms in a way that scales horizontally.