Engineering capacity on contract, alongside a team that already exists or as the resource a project is short of. By the day or by the engagement, set out in writing the same way a fixed-scope build is.
Ask about contractingMost often a model that works in a notebook and has stalled on the way to production, or a delivery date that needs a second pair of hands rather than a second opinion. Both are capacity problems, and both are quoted by time rather than by scope.
The engineering underneath an AI system is most of the job. The model is rarely the hard part.
Capacity, not a project. The scope stays with whoever owns the roadmap.
Ten years building and running AI and machine-learning systems, a Masters in Computer Science, and delivery experience across ecommerce, industrial equipment and professional services — which means arriving already familiar with how a delivery team works, rather than learning it on the engagement.
What the project is, what it is short of, and roughly how long for. That is usually enough to say whether it is a fit.