Outside Spy
Glasgow, UK
This business has never run machine learning in production. This is the role that changes that, and it starts on the single highest-value optimisation problem they have.
It's a physical manufacturing process where getting the settings right first time, instead of running it again, saves material, energy and time at very large scale. Today that judgement sits with a handful of specialists. You'd build the ML capability that supports and extends it, from data pipeline and training workflow through to serving, monitoring and the standards every model that follows will be held to.
You'll begin with no ML platform, no playbook, no team to slot into. You define all of it. The data involved is also the company's most sensitive IP, so everything runs inside a closed perimeter with no external egress, and that constraint shapes how you build from day one.
What this needs:
You're senior enough to be the last word on ML decisions here, because...