Automation built lean, tested against real data, and running right now — not a mockup.
Before automation, I spent years in roles where a single missed detail meant a real problem — verifying property and maintenance records, processing official government documents, reviewing data for compliance. That background shaped how I build systems today: precise, tested against edge cases, and built so nothing quietly slips through.
I'm currently deep in AI automation — Zapier, MCP, n8n, and AI agent design — bringing that same reliability into workflows instead of spreadsheets. If a process is manual, repetitive, or held together by memory, I build it a system instead.
Click any tool for a quick note on how I use it.
Each one runs against real accounts, real APIs, and real data — not slides.
A website contact form logs new leads instantly, alerts the team, checks for duplicates before saving, and automatically flags any lead with no follow-up after 48 hours — so nothing goes cold silently.
A live-calendar booking system, built to remove the back-and-forth of scheduling a consultation. Here's exactly how it works and where it stands right now.
How it works: a visitor requests to book a time, and the system checks Google Calendar in real time to see what's genuinely free over the next 7 days — no fake availability, no double-booking. The moment a slot is picked, it re-checks that exact slot one more time (to catch the rare case where two people try to book the same moment seconds apart), then creates a real calendar event, logs the booking to a spreadsheet, and fires off both a confirmation email to the client and a notification to me. A separate scheduled process checks hourly for any booking happening in ~24 hours and sends a reminder automatically — once, never twice, for the same booking.
What's already working: real-time availability checks, duplicate/double-booking prevention, calendar event creation, CRM logging, confirmation + owner emails, and the 24-hour reminder watcher — all tested against a real Google Calendar and real email accounts, not mocked data.
What I'm still refining: the booking page's visual calendar display (moving from a slot list to a full month-view grid), and anchoring the time-zone logic to a specific client's actual location rather than a fixed default — the kind of detail that's best finalized against a real client's real timezone rather than guessed at in advance.
Maintained and verified large volumes of property records, reviewed documentation for compliance, and identified inconsistencies before they became problems.
Processed and verified official government documents with strict accuracy, escalating anything incomplete rather than letting it pass through.
Content evaluation, classification, and quality review for AI training data — hands-on exposure to how AI systems are built and judged for accuracy.
Six certifications covering Zapier automation and MCP, from foundations through AI agent design.