Facilities maintenance by trade. Self-taught builder by curiosity.

I’m not a conventional developer, and I didn’t take a conventional path into software.

I build things because building is how I learn.

Most of what I know has come from finding a question I couldn't leave alone, trying to make something around it, discovering everything I didn't understand, and following that trail until I understood a little more.

Sometimes that produces useful software.

Sometimes it produces an experiment that teaches me why an idea doesn't work.

Both count.

My projects tend to wander around the edges of:

• AI systems and agent workflows
• persistent memory and continuity
• local-first software
• human control and system authority
• persistent characters and human-computer interaction
• tools for building with AI without handing the whole process over to it

I’m especially interested in places where the technical question eventually becomes a human one.

Who owns the memory?

Where does authority live?

What should remain deterministic?

What happens when the model changes?

How do we know what is actually true about the system we're building?

Those questions have led me into some strange places. I like that.

A few of the things currently occupying my brain:

Addison An evolving exploration of operator-owned continuity, memory, and state for working with AI systems.

Veyra A persistent character runtime exploring the boundary between cognition, continuity, portrayal, and embodiment.

WBTSR A human-governed way of building with AI agents through bounded work, explicit project truth, verification, and rehydration.

There are smaller experiments too. Some will become real projects. Some will teach me something and disappear.

That's part of the process.

I used to put more pressure on my work to become an opportunity.

I'm trying to think about it differently.

I build to learn.

If something I make eventually creates an opportunity, finds users, helps somebody, or turns into something larger, that's great.

But it doesn't have to.

The work has already changed the way I understand software, systems, AI, and my own ability to learn difficult things.

I'm interested in continuing that process and becoming capable of doing unusual things.

I use AI heavily in my development process, but I don't believe using AI means giving up human judgment or responsibility.

I try to:

• understand what I'm building rather than only produce it
• keep important decisions and system truth inspectable
• separate proposals from authorized actions
• document failures and direction changes instead of hiding them
• change my mind when the evidence earns it
• leave myself able to understand the project when I return later

I'm still learning.

That's not a disclaimer. That's the point.

This GitHub is less a portfolio than a trail of things I'm learning to build.

Some repositories are polished enough to share.

Some are experiments.

Some are snapshots of ideas that have already evolved beyond them.

If something here is useful or interesting to you, feel free to explore.

• NQL Systems: https://github.com/NQL-Systems/
• Website: https://zerolxgic.github.io/
• X: https://x.com/ZerriusVale