Projects

Built and running.

Real systems in production — each one designed, built, and shipped end to end by me.

AI BOARDROOM · WEB APP

Executive Command Center

The problem: big decisions get made with one perspective in the room and no paper trail.

What I built: a live production web app where a team of AI agents debates your real business decisions across multiple rounds and searches the web for current facts. The agents can be configured as any roles on a team of any size — the five-executive boardroom shown here is just one setup.

Running now: you run the whole room by voice or text, every verdict is logged, and any discussion exports straight to Word, Excel, or PowerPoint.

Recovery Assistant mobile app home screen in a phone frame, showing a daily readiness ring, quick stats, pain logging, and today's checklist

AI HEALTH COMPANION · MOBILE APP

Recovery Assistant

The problem: long recoveries fail in the boring middle — when every day feels the same and the plan stops adapting.

What I built: a native mobile app with an autonomous AI agent at its core. You log a few data points a day; every morning the agent reviews them, adapts the day’s plan, and delivers it to your phone.

Running now: one system end to end — mobile app, web dashboard, and a cloud backend that runs its daily check entirely on its own.

AZURE OPENAI · NL-TO-SQL

Azure AI Chatbot for SQL Server

The problem: the answers are in the database, but getting them means waiting on whoever writes SQL.

What I built: a web chatbot backed by Azure OpenAI that turns a plain-English question into T-SQL, runs it against a live SQL Server database, and returns the results in a table — with the generated query and a step-by-step explanation alongside.

Running now: it handles real analytical asks — top-N rankings, median calculations with CTEs — without anyone writing a line of SQL.

AI DATA PIPELINE · AZURE

Azure AI Chatbot + SQL Server + Power Automate

The problem: even a good answer still has to be copied, formatted, and emailed to the people who asked.

What I built: an extension of the SQL chatbot where a Power Automate flow emails the full result set to your inbox as a formatted table the moment the answer comes back.

Running now: ask a question in plain English; the report arrives without another click.

CLAUDE + MAKE · AUTOMATION

Daily News Digest

The problem: staying current takes an hour a day you don't have.

What I built: a fully cloud-based automation — Claude generates two curated news digests (one on crypto investing, one on AI) as formatted HTML with clickable article links, and a Make scenario schedules the run and sends the emails.

Running now: both digests land in a Gmail inbox at 7:00 AM every morning, with zero manual effort.

CLIENT WEBSITE · LIVE

Kidlanta

The problem: a metro-Atlanta kids-events business needs parents to find events, attractions, and shows in one place.

What I built: a multi-page website designed and built from the owner's own sketches and content plan — four browsable sections, a this-weekend events rail, live audience filters, per-event detail with official links and directions, and a newsletter signup — plus an agent-run content pipeline: a scheduled agent researches real metro-Atlanta kid events every week and sends the owner one approval email, and approved events publish to the site automatically.

Running now: live at kidlanta.com — listings stay current through the weekly research-and-approve loop, and the owner adds or changes content by sending a plain-English email.

Next

Want a system like this running in your business?

Tell me what you're trying to automate. I scope it, build it, and run it.