About
I connect operations experience with practical AI systems.
My work spans service operations, infrastructure, analytics, documentation, automation, and AI-enabled workflows. The common thread is making complex work more observable and easier to operate.
I enjoy the layer below the demo: identity, permissions, data boundaries, testing, evidence, monitoring, recovery, and handoff. Those details determine whether an interesting prototype becomes a useful system—or a future support problem.
I am building in public carefully. Some projects are working prototypes, some are private operating systems, and some are design methods still being tested. I label those states because the distinction matters.
I am based in Canada and interested in conversations and opportunities globally involving AI systems, operations, automation, technical program work, service design, analytics, or evidence-rich documentation.
How I decide what to publish
Every article on this site carries an evidence state, and it is the honest one. Live means I observed it. Built means it is implemented but not necessarily running now. Historical means a dated incident with surviving evidence. Design means proposed and not yet delivered.
Separately, some material is safe to publish only in generalized form. When the underlying system is a live secret store, an access path, or anything else where the specifics would help an attacker more than they help a reader, I publish the pattern and keep the implementation private. Articles written that way say so at the top.
The full standard is written up here, including the release checklist I use before anything goes public.