Making software got easy. Making software worth having didn't.

01 · The accomplishment

I didn't learn to code. I built a company.

CircleRootTech is a technology company I am building from the ground up on one conviction: software should be deliberate, it should be durable, and it should be made for the people who use it rather than at their expense. That's the whole brief. Not scale, not an exit, not shipping fast enough to flood a store with things nobody asked for.

The products are the output. The company is the thing I actually made: registered in Utah, with an operating handbook, a strategic roadmap, a pricing model, and a standard for what ships — run by one person, alongside a full-time job.

Learning React Native was the easy part. The hard part was deciding what a company is for. There is an enormous amount of software being generated right now by people who can prompt their way to a build, can't get it through review, can't support it, and never asked whether it should exist. I'm not a vibe coder shipping volume. I don't compete on how much I can put out. I compete on whether the thing is worth having.

Money isn't the goal. Revenue keeps the company alive so it can keep improving — that is its entire job. What I'm actually building for is access: that the people a product was made for can afford it, can trust it, and get a better version of it every year they stay.

The charter

• No ads, no data selling, ever. Not a growth-stage decision to revisit later. It's the reason the company exists.
• Priced for access, not extraction. If the choice is between a bigger number and more people being able to use the thing, more people wins. Every time.
• Every product comes from something I've lived. Community, safety, caregiving, connection, getting hired. I don't build for markets I've only read about.
• A three-year improvement cadence. Apps get better on a schedule, driven by user feedback — not abandoned at v1.0, not churned for the sake of a changelog.
• Status is honest. One app is live. The rest are in development, in review, or being rebuilt — and this page says so instead of rounding up.
• I test every fix myself before it ships. On my own device, every time.

04 · Newest product

ApplyWorthy

Job seekers spend hours on applications that were never going to be read — knocked out by an applicant tracking system before a human sees them. ApplyWorthy checks the friction before you apply, so you spend your hours where they can land.

I built it because I watched it happen for sixteen years, from the side of the desk where you have to tell someone why they never heard back. It ships as a licensed tool for workforce-development organizations — the people already doing this work, given something that scales past one caseworker at a time.

Node · ATS detection engine · profile-matched scoring · B2B2C licensing model

See the full ApplyWorthy page →

06 · Working with AI

I'm not an AI researcher. I'm someone who ships with it.

Every app in this portfolio was built by one person with AI as a working partner — architecture, code review, debugging, strategy. That's not a novelty anymore. What's still rare is knowing where the partnership ends.

AI will confidently help you build something that cannot work. It helped me build a child-safety app around a capability iOS was never going to grant me — and no amount of prompting would have caught that. A person had to. That experience is most of what I know about working with these tools: they compress the building, and they raise the price of bad judgment. The judgment is still the job.

It also turned into a service. There is a flood of AI-generated software right now that can't pass App Store review, and the people who generated it often can't read it well enough to find out why. I can — because I've been rejected for most of the reasons already.