Building in Public

Published: 2026-02-24 · 3 min read

Most people doing this kind of work don't show their work. I am.

Ridley Research is where I document what I'm building — AI agents that actually work in production. Not demos. Not pilots. Agents that run overnight, write memory, monitor systems, and surface what matters.

Everything on this site comes from a live system running on a Mac mini in Plano, Texas — handling security audits, research digests, writing drafts, memory consolidation, and now this blog post.

That's the point of this site. I'm documenting what I build, what breaks, and what I learn while building it. The lessons here are real because they cost me something — tokens, time, or a late-night session figuring out why a permissions misconfiguration was blocking deploys at midnight.

What I'm building

The core product is a replicable AI infrastructure stack for small teams. One agent as chief of staff. Specialized sub-agents for security, research, content, and ops. Memory that persists across sessions. Crons that run while you sleep and brief you in the morning.

I call it the Enoch stack. I pressure-tested it on my own operations first.

Why document it

Because the gap between "AI is amazing" and "AI is running my operations" is enormous, and almost nobody is publishing honest content about what it actually takes to cross it. I am.

Every post here is drawn from something that actually happened in the build. If a cron job failed fourteen times before I fixed the allowlist, that's a post. If I learned the hard way that model swaps from chat can take down the gateway — that's a post too.

Subscribe or check back. I'll keep shipping.

— Ridley Research, Ridley Research