Scope

This role owns the paved road: data access, training and evaluation tooling, deployment, and the cost of running models. Product teams should not each invent their own stack.

Required experience

Has run platform engineering for data or machine learning. Has had internal customers, a backlog, and a bill. Can say what the company should buy and what it should run itself.

Interview focus

Ask who the customers of the platform were, what those teams stopped building themselves, and how cost was shown to the CTO.

Sample job description

Report to the CTO. Run the platform AI teams build on. Set the standards for data access, evaluation, and deployment. Keep the cost of inference and training visible. Hire the engineers who make that path the default.