Engineering decisions we can examine together.
I teach at Sheridan and Conestoga and work on software systems. These are proposed session topics for developers and engineering leaders; format and availability can be discussed.
Reviewing an AI-assisted change
Work through a shipment mapper that passes its existing fixtures while introducing an unsupported default. Examine what the tests establish, identify the missing case, and explain a review decision.
Where an agent's authority ends
Explore a browser-agent design in which a model classifies an action and application code decides whether it may run. Discuss approval, changing state, and what an evidence log can establish.
Read the revision-bound approval article, then use the team evaluation guide to examine one boundary. The experimental Stopline repository contains the implementation.
Measuring a local-first coding agent
Use a small command-line agent as a case study in boundary design: what the tool can observe, which actions it may propose, and what evidence remains after a run.
For a private team session applying the review exercise to a real process, see the workshop interest page.
To discuss a session, contact me on LinkedIn or email me.