Software architect focused on enterprise systems and AI-assisted development workflows.
I'm interested in what stays durable as coding agents get stronger — preserving intent, human decision authority, verifiable project state, and evidence across long-running AI-assisted work, so it remains reviewable and recoverable.
Currently building ai-workflow-harness and regularly using it across my own projects.
I write about these experiments — agent collaboration, review workflows, and what I learn along the way — on my blog (in Korean and English).




