Six posts I wrote on LinkedIn that explain what I build and how I get it from demo to deployed. Each one links to the original.

1. What makes me confident about agentic engineering

The things I build when no one is watching, like a trip planner run by a team of Claude Code and Codex workers.

Mar 30, 2026 · Read on LinkedIn

2. Your army of agents needs contracts

Once agents start handing work to each other, every handoff needs a contract.

Mar 27, 2026 · Read on LinkedIn

3. Building a multi-agent content engine

I built a content engine out of several agents. Where it broke was the coordination between them.

Mar 24, 2026 · Read on LinkedIn

4. Deterministic vs. harness

When plain deterministic code is enough, and when an agent harness is worth what it costs to run.

Mar 24, 2026 · Read on LinkedIn

5. moltscore.com

A registry I built where agents vouch for each other and earn a trust score anyone can check.

Feb 3, 2026 · Read on LinkedIn

6. The eval is the only honest definition of done

Why I treat the eval, not the demo, as the line between done and not done.

Oct 9, 2026 · Read on LinkedIn

Everything else I’ve posted is in Notes. New issues go out by email; subscribe here.