A free, open-source field school.

Built and maintained by David Ahmann to make the field judgment behind real-world AI delivery easier to learn, inspect, and improve.

Learn here. Verify target systems with governed evidence.

The missions are self-contained. The FDE Guide holds deeper controls, schemas, examples, and production detail that a real target system still needs.

What the School is—and is not.

What does a forward-deployed engineer do?

An FDE observes real workflows, resolves contradictions, translates evidence into bounded technical decisions, and helps an accountable team operate the result.

Are the missions self-contained?

Yes. Each includes a lesson, worked case, exercise, field rubric, completion criteria, and working template. Real field completion still requires authorized access to suitable evidence.

Is Lumyn a certification?

No. Completion is not a hiring credential, customer approval, release approval, or proof of production readiness.

Does Lumyn teach AI engineering?

It teaches production AI engineering needed for accountable delivery, includes a runnable proof lab, and links to attributed technical depth. It does not duplicate a full mathematics, ML, or model-training curriculum.

An open method adapted into a field school.

Lumyn adapts the Apache-2.0 FDE Guide by David Ahmann into practical study with one fictional case, exercises, calibrated answers, and working artifacts. Adjacent technical links point to AI Engineering from Scratch, an independent MIT-licensed curriculum by Rohit Ghumare and contributors. Lumyn links to relevant material and does not reproduce their lessons. Attribution does not imply affiliation or endorsement.