AI Engineering project charter

Purpose

Define the practical problem this project will solve. Do not assume that “AI Engineering” already names a distinct discipline; that is a research question, not a starting fact.

Intended users

Not yet established.

First bounded outcome

Not yet selected. Choose one outcome that can be delivered and evaluated in two to four weeks.

Included

• Definition of the first user and communication problem.
• A working vertical slice.
• Evidence and decision traceability.
• Evaluation of the Repository Operating System pilot.

Excluded

• Broad discipline claims without comparative evidence.
• An exhaustive communication taxonomy.
• Autonomous acceptance of research or policy.
• Production handling of secrets or sensitive communication data before a privacy and threat review.

Success criteria

• The first vertical slice has observable acceptance tests.
• Material decisions cite their evidence and alternatives.
• A successor can continue from repository records without chat history.
• Pilot measurements can compare the operating approach with a declared lightweight baseline.

Constraints and assumptions

• Constraints: not yet established.
• Assumption: a bounded communication problem can be selected without first resolving the full disciplinary boundary.

Owners and decision authority

Not yet assigned.