AI as engineering leverage
Daily work with local and hosted LLMs, coding agents, planning loops, and review checkpoints — designed so the human keeps judgment and the agent keeps the typing.
A workshop for LLM-assisted systems, built by someone who's shipped production telemetry. This is my public home for experiments, projects, and field notes around agentic coding workflows, LLM tooling, and the infrastructure that keeps them honest. Industrial software background is there — SCADA, telemetry, edge + cloud — but the page leads with the work, not the résumé.
One anchor: agents as a daily engineering tool. One support: the infrastructure habits that keep them trustworthy.
Daily work with local and hosted LLMs, coding agents, planning loops, and review checkpoints — designed so the human keeps judgment and the agent keeps the typing.
Production habits give agent work somewhere safe to land: observability, boundaries, rollback thinking.
Each card names what was built and where agents materially helped — honestly, not inflated.
This site: an AI-first personal workshop with Astro, Cloudflare Pages deployment, real legal routes, and a deliberately static privacy posture.
An Astro + Cloudflare site for a fiction project, including newsletter backend work and content automation patterns.
Enough to show thinking, not enough to become a content treadmill.
Freshness checks, authorization, observability and command safety as transferable habits.
Coming soon. Where automation helps, where it needs boundaries, and what stays human-owned.
Coming soon. How to show technical signal without publishing operational internals.
Only intentionally public destinations belong here.
Short notes and selected public experiments appear here as they become worth sharing.
The public contact surface stays small and deliberate.