I build the version that still works later.

I got into computer science because I wanted to know how things actually worked, and software was the only field where I could open the thing up and look. That turned into a degree with an AI concentration at UNC Wilmington, back when the interesting question was still whether a model could do a task at all.

Now it can do most of them, badly, on the first try. The interesting question moved. It is no longer whether a model can do the work — it is whether you can trust what it did, explain it to someone else, and switch it off when it goes wrong.

GrizzlyBare LLC is my practice, based in Charlotte, NC. I build AI agents, internal applications, and websites — written by hand, not assembled from a builder.

The through-line is that everything ships with its seams visible: scoped permissions, an activity log you can read, a set of evaluations that say what was tested, and a documented way to turn it off. Not because it is best practice, but because it is the difference between a system you own and a system that owns you.

We start with a call about what you are trying to do, not what you want built. Those are different, and the gap is usually where the useful conversation is. I will tell you if the thing you asked for is more than you need — I have talked people out of an agent and into something smaller more than once.

Then I scope it, build it, and give you the keys — the code, the accounts, all of it. You own it outright. There is no monthly fee holding it hostage, and no platform you have to keep paying to keep your own work.