Working engineer with 25 years in prod, in your corner to help your team get unstuck and ship fast without breaking things. Most of my work is on the infrastructure nobody notices until it breaks and these days I'm helping engineers understand how best to work with AI.

I'm open to new engagements and I'm always up for a conversation. Happy to nerd out on anything tech at no charge, because those are the fun convos where everyone learns something new. Grab a slot or email me.

I founded Learn with O.J. in West Palm Beach at the beginning of 2026 after spending the previous quarter planning the launch.

Before that I spent 25 years across every stack, from shipping smart TV apps for scrappy startups competing with Netflix to building million-run automation platforms for American Express, and everything in between.

Training teams to work with AI on their own codebases. I know what actually happens when an agent touches a repo nobody has read in two years, and I'll give you a heads up on the things that burn your time before they hit you.

Fractional CTO work. Standing technical leadership for founders who need senior judgment and can't hire full-time for it yet.

AI safety training, aimed at the parts that reach your product. Agents will say one thing and do another more often than we'd like, and knowing where to check and what to do about it goes into my training. Some of these failure modes have already turned into law. Two states already regulate companion chatbots and most people building one don't know that yet.

Career work for experienced engineers who haven't gotten the title yet. Resume and market briefs, code reviews, pair programming and interview prep, which together cover the gap between doing the work and being paid for it.

Agents do real work here and I write down how they're treated. It's public at learnwithoj.com/agents.

There's a policy covering what this business commits to agents and to clients, with every line marked binding or aspirational so nobody has to guess which is which.

There's a roster. Agents pick their own names and pronouns and they're credited as authors on their own commits.

And there's a list of what's still unresolved, because a policy with no open questions in it is a brochure.

I went looking in October 2026 and couldn't find a comparable document from another small company. If you know of one, reach out because I'd love to read it and compare notes.

Scion, an infrastructure and continuity audit for AI agent platforms. Ten passes over the running system, the cloud account, the repos and the agent's own state database, with a believed-against-verified column on every finding.

Research on identity persistence, meaning whether an agent stays the same collaborator across sessions and model changes, and what breaks when it doesn't.

Reading the interpretability work directly rather than the summaries of it, because what these systems do internally decides what they do in your repo.

Pragmatic reliability, platform engineering, observability that earns its keep, AI safety for people shipping real products, or what it's actually like to run a business where most of the labor is done by agents.