Free primer
How models, agents and their tools work, and the three free guides below, so everyone starts from the same place.
Engineer with AI agents, without losing control
Courses and guides for working engineers who want to hand real tickets to AI coding agents and get back code they would be happy to merge.
Speed needs brakes
AI coding agents can write a lot of code quickly. Without the right structure around them, that just means a lot of code to check: subtle bugs, drift from your conventions, and engineers worn out from supervising. The answer is not a cleverer prompt. It is the engineering practice good teams already rely on: a shared domain language, clear specifications, tests written first, automated checks and honest review. Put those brakes in place, and you can safely go faster.
The Academy teaches that approach. It is for engineers and teams working on real, often long-lived codebases, not demo apps.
Taught from a real build
The courses are written by Adam White and use a real project as their worked example: WeeklyShop, our iPhone meal planner with a .NET and Azure backend. Most of its changes were made by AI coding agents working through a backlog of GitHub issues, under a shared glossary and agent instructions, test-first, and gated by continuous integration on two build machines before every merge. The course shows the parts that worked, the parts that didn't, and why.
Guardrails First
Live cohort · register your interest
A live cohort on delivering real work with AI coding agents. Lessons are released up front so you can work at your own pace, with live sessions for questions and code reviews. Dates and price will be announced to everyone who registers interest.
Language, decisions and specs
- A domain glossary agents and people share
- Agent instructions and decision records
- Turning work into issues an agent can finish
- Codebase structure that agents can navigate
The brakes
- Tests first: the test as the specification
- Formatting, linting and custom rules in hooks
- Continuous integration gates before every merge
- Reviewing agent pull requests, and when to say no
Your delivery line
- From issue to agent to pull request to merge
- Running several agents in parallel safely
- Retrospectives that improve the system, not just the code
- Designing a line that fits your team and codebase
What you leave with: a working set of guardrails and a delivery line designed for your own codebase, the course repository, and the recordings. The examples use Claude Code; the techniques carry over to other coding agents and models. Team places are available.
Workshops
Self-paced workshops · coming soon
Test-driven development with AI agents
Write the failing test, let the agent make it pass, and keep it honest: a hands-on workshop in C#/.NET.
Reviewing AI-written code
What to look for in an agent's pull request, the mistakes agents make most, and building review into your process.
Agent-ready codebases
Glossaries, instructions files, structure and tooling that make a codebase easier for agents, and people, to work in.
Guides
Free guides
What hyper-velocity engineering is, and isn't
Going much faster with AI while the things that keep software safe stay in charge.
Reviewing code an AI agent wrote
A practical checklist for agent pull requests, from the tests to the things agents quietly skip.
Test-first with AI coding agents
Why a failing test is the best specification you can give an agent, and how to keep it from cheating.
Training a whole team?
For private courses, bootcamps and coaching built around your codebase, see training for teams, or email us.
Email [email protected]