AI Lab content systems
Mar to Sep 2026AI tutorials for a nine-figure education company, on a pipeline I built.
200+ videos
teaching AI to non-technical adults

What I built
Idea engine
New video ideas are checked against everything already made, researched first, then scored on a rubric with hard gates, so only strong ideas reach a script.
Publishing in one command
Upload into the right Wistia series, a loudness check before members hear it, the tracker updated and the thumbnail set. A newsletter builder stages each send for a person to approve.
Across models
I taught Grok on the X API alongside Claude and OpenAI, and built the shared context so Claude Code and Codex worked from the same rules.
“You have the double skill set, which is the AI itself plus presenting, and that is more valuable than just the AI.”
Emils Pakarklis, Founder, iPhone Photography School
Engineering
Models draft, a person publishes.
- How it fits together
- Shared context files for Claude Code and Codex, feeding scripts, publishing, newsletters and a weekly KPI report.
- Tested
- A lint on the KPI report.
- Claude Code
- Codex
- MCP
- Wistia API
- X API
My role. I built it as the in-house Applied AI Engineer.