The title is new enough that every organisation defines it differently. Underneath the variation, the role has a stable core: translating business intent into an AI system design that engineers can build, operators can run and the business can trust. This lesson draws that boundary.
Between strategy and implementation
Leadership decides that AI matters. Engineers build models and services. The architect sits between them, turning a goal into a design with defined components, interfaces, data flows, risks and costs, and defending that design as it meets reality.
The decisions that are yours
Build versus buy versus fine-tune. Which model class for which problem. Where data lives and how it moves. What the serving platform looks like. How quality, safety and cost are measured. Who is accountable for each.
The decisions that are not
Hyperparameters, code structure and individual tool choices belong to the engineers. An architect who makes those becomes a bottleneck and loses the trust of the team.
Action Step
Write a one-page role charter for an AI architect in an organisation you know: the decisions owned, the decisions delegated, and who the role reports to and works with.
This course is vendor-independent: it is not affiliated with, endorsed by or accredited by any tool vendor or certification body, names products only for identification, and issues no credential. Verify current documentation before applying anything in production.