I measure things, then I build the thing that fixes them.

Construction work zone safety software. Safety representatives author Internal Traffic Control Plans on a canvas, placing equipment and worker zones, and the app flags struck-by conflicts before anyone sets foot on the road. Plans publish to a QR code so the crew acknowledges them on a phone at the job site.

Struck-by incidents are a leading cause of construction fatalities, and the plans meant to prevent them are still drawn in marker on a tailgate. The target is a safety rep authoring and publishing a real plan in under thirty minutes.

Two products aimed at opposite ends of the pest control industry. Bugsy is a free consumer app that identifies pests from a photo of droppings, tracks or damage using image recognition, returns a confidence score, and hands back either a DIY treatment path or a route to a qualified local pro. The PestTech Management Portal is the other half: a cloud back-of-house for pest control companies covering scheduling, customer chat, marketing automation, analytics and payments.

Co-founder and technical lead on a four person founding team, alongside a pest control operator, a marketing director and a business development lead. Initial markets Oklahoma, Colorado and Pennsylvania. The thesis is simple: consumers guess wrong before they call anyone, and the companies they call are running on fragmented tools.

An AI powered paper review tool for educators. Teachers upload student writing and get structured, rubric aware feedback back, cutting the grading pass from hours to minutes while keeping the teacher in the loop on every judgment.

A full production stack across three services, built and debugged solo. It has since been assessed as a sellable asset on acquisition marketplaces.

112 browser based calculators and utilities across 8 categories, roughly 170 URLs. Every tool runs entirely client side: no uploads, no accounts, no display advertising.

An experiment I am running in public. Impressions jumped roughly a hundredfold after a crawlability fix while clicks stayed flat, which is exactly the trap most content sites fall into. I rebuilt one calculator around the real search intent behind its impressions and I am judging the whole strategy on whether that single page moves. If it does not, I will say so.

A native iOS app that put live flight status and destination weather in one place: schedules, delays, cancellations and gate changes next to current conditions and a five day forecast, with push alerts when either one changed.

My first end to end product. I wrote the white paper, the competitive analysis, the tiered pricing model and the feature roadmap, then built and shipped it. Most of what I know about scoping a product I learned from the parts of this one I got wrong.

A grid based 3D base builder for iPhone where you run an orbital debris cleanup business. Launch pads, barracks, hangars: the base is the navigation, and each building opens its own function. Fallout Shelter crossed with SimCity, in orbit.

Adapted from my father's book, Space Debris: Clearing the Final Frontier. He is my first TestFlight tester, which is both a quality bar and a good reason to keep the release branch clean.

An incremental management game where you build a pest control company from one guy in a van into an empire. Fallout Shelter style: hire technicians, upgrade the depot, take contracts, watch the operation grow while you are not looking.

This one came out of PestTech. Spending months inside the real industry gave me the domain detail a pest control game actually needs, and building the game clarified the operational loop the real product has to model. Each one makes the other sharper.

An AI generated desert art brand with an audience of roughly 11,000 and a print on demand storefront selling canvases, shirts and color change mugs. Behind it is a five stage content pipeline: generate, curate, produce, stage, report, with a hard approval gate before anything publishes.

I grew the original account from zero to about 1,800 followers by hand, then lost it to an automation policy strike I caused myself. The rebuild kept automation on production and took it fully off engagement. Expensive lesson, worth it.

A local AI agent stack on my own hardware: a supervisor agent with an Obsidian vault as its workspace, a local inference gateway, and Claude Code as the execution layer for filesystem and shell work. Alongside it, a four agent Python orchestration system for market analysis with a hard simulation mode safety flag.

I wanted a first hand read on where small local models break down on tool calling rather than taking anyone's word for it. They break down early, and knowing exactly where changes what you hand to a frontier model.