Field Notes / ai-engineering
AI engineering field notes
Reports about putting AI into engineering work without losing review, provenance, or technical ownership.
Scope boundary: This hub covers engineering control. Product safety remains in its own domain.
Operating situation
Inspect the problem: AI Engineering ControlInspectable proof
System Digital Twin in Four Weeks
How two people built a textual system twin for a large inherited codebase in four weeks.
Read the case: System Digital Twin in Four WeeksEvery note in this hub
This list includes published English notes only.
AI Engineering Control
Two roadmaps make fresh code feel like legacy
What a repository reveals when the team's real planning horizon is shorter than the official product plan.
Two roadmaps make fresh code feel like legacyAI Engineering Control
Agents work better as advisors
A practical boundary for agentic engineering when confident automation is still wrong often enough to matter.
Agents work better as advisorsAI Engineering Control
Choose the first AI workflow worth changing
A practical way to find one costly delivery workflow, test whether AI belongs in it, and define a pilot without starting a transformation program.
Choose the first AI workflow worth changingAI Engineering Control
Pilot AI on one real software delivery task
A bounded pilot method for testing AI in software development with a baseline, human review, stop conditions, and a clear scale-or-stop decision.
Pilot AI on one real software delivery task