Product AI Engineer
Ship the feature users hit. Leave someone who can extend it.
For product and engineering leaders who already have a product and need an AI feature in the path users hit. This seat ships that feature, the eval that decides good enough, and a teammate who can extend it.
We have a product. I need the AI feature in it, and someone here who can change it next month.
What the Product AI Engineer owns
The path users hit
The production path, with the model inside it. Not a side demo.
Good enough
The eval that decides whether the feature stays on, written before the feature stays on.
A teammate
Someone on your side who can extend that path after the milestone ends.
This is the person who builds. The spec seat is AI Product Manager: the task, the escalation path, and the kill criteria. A feature PM writing tickets, with nobody building, is a different engagement.
Where this seat sits
AI Pivot Squad
Primary home. The squad is the team. This seat is the person who builds the feature.
AI Product Manager
The spec seat. This page is the person who builds.
Forward Deployment Engineer
Operators, and one workflow on the system they already run.
Common questions
No. AI Product Manager specs the outcome. This seat builds the feature on the path users hit. Use the spec seat when the team cannot yet say what good enough is.
No. The product already exists. This seat puts a model on a path users already take. A blank-page product is a different engagement.
No. The leave-behind is a feature users hit, an eval, and a teammate who can change it. Evals, traces, and a failure budget around a demo that is not yet a feature are the Harness Engineer.
Put the model on the path users hit.
A 30-minute scoping call is enough to tell whether this seat is the entry, or whether the squad should come with it.