I build the systems behind a live trade show, the AI tools our IT team uses every day, and the workflow that keeps this site accountable. These three projects show how I plan, integrate and document work that other people depend on.
01 · Live-event infrastructure
American Film Market IT operation
The American Film Market turns a hotel into an international trade show: guest rooms become staff and exhibitor offices, with screenings in a nearby theater. I have led its IT operation for 23 consecutive years. The deadline stays fixed, even when venues change and on-site time gets shorter.
My role
As sole architect, planner and on-site lead, I own the inventory, timeline, staffing, vendors and runbook. I coordinate with venue IT, direct a team of three, lead the office-to-hotel cutover and run the systems throughout the show.
Tools and approach
I design the network around hotel and theater constraints, stage rental equipment, and author the annual golden laptop image for the vendor to clone. The runbook ties the technical build to the people, equipment and sequence needed on site.
Why it matters: The cloned image lets each user deploy in minutes. Planning and documentation carry more of the load as a smaller crew delivers the same build with less preparation time.
02 · Team automation
AI tooling for the MSP team
At our managed service provider, investigating an alert can span ticketing, remote monitoring, endpoint security, Microsoft 365 and backups. I build the connections that let Claude work across that stack with structured, scoped access.
My role
I led the team’s Claude rollout and built the Model Context Protocol (MCP) servers, agent skills and integrations behind it. The ticket-triage skill investigates alerts across systems, writes a private note and closes informational tickets.
Tools and approach
My MCP servers connect Microsoft 365, N-able N-sight, WatchGuard EPDR and Freshdesk. Most are read-only; M365 administration previews changes before execution. An n8n pipeline handles reported phishing email, while a weekly backup review gathers alerts and flags jobs for attention.
Why it matters: The team can use AI for investigation, documentation and recurring reviews within its existing tools. Ticket triage is our most-used skill, and the backup report gives us specific jobs to examine.
03 · Reversible publishing
This AI-managed WordPress site
This site is both my personal site and a working AI lab. With several agents making changes, I need a shared process that records who did what and gives me a way to restore an earlier version.
My role
I use Claude, Codex and OpenCode to manage the site under one set of harness guidelines. The workflow takes a snapshot before a write, verifies the result, exports and commits the change, and closes the session with a numbered Lab Note.
Tools and approach
WordPress runs in Docker on my Linux host, reachable over Tailscale. Agents use SSH, WP-CLI and the WordPress MCP Adapter. GitHub stores readable content exports and restorable site data; rollback scripts can restore a snapshot or an earlier commit.
Why it matters: A change, commit, rollback and verification round trip has been tested. The Lab Notes make the experiments visible, including what worked, what failed and how to undo a session.