Four standalone courses on how AI actually works — and how to build with it. Buy one. Buy all four. Watch at 2× at 11pm. Discord community included.
First principles of large language models — tokens, transformers, training, hallucinations. The mental model that makes everything else click.
You stop guessing how the model works. You understand it well enough to trust your own instincts about it.
Stop guessing. The Prompting Framework, chain-of-thought reasoning, and patterns that make models do what you actually want.
You stop "trying prompts until one works." You write prompts that work the first time — and you know which lever to pull when they don't.
Build with AI coding agents. The risk/capability matrix, prompt injection, the agent loop, and CLAUDE.md context architecture.
You ship a working AI agent — and you understand what it's doing well enough to debug it, secure it, and extend it.
Make your AI app smart, connected, and automatic. Claude /skills, Supabase, edge functions, and triggers that run on autopilot.
You go from "I built a thing" to "my thing runs without me." Apps that remember, react, and fire on their own.
I built AI/ML systems at Meta and led AI/ML at Brev.dev (acquired by NVIDIA in 2024), with a B.S. and M.S. in Computer Science from Stanford specializing in AI — where I was a TA for PhD-level core AI courses with consistently excellent feedback. Now I teach AI on Instagram, YouTube, and right here.
I built these courses because the gap I see most often isn't capability — it's confidence. People who could build don't, because nobody walked them through it. So that's what we do here: walk through it, end-to-end, no hand-waving.
Buying any course unlocks the Harper Carroll AI Discord — a private space for builders working through the same material. Channels per course, async office hours, and a weekly thread where I drop in to answer questions.
It's like having a study group, on your own schedule.
↳ questions? hai@harpercarrollai.com