Skip to content
Course · coming soon

Learn to build with AI.

Software engineering principles for people who build with AI, explained so you know why, not just how.

The AI writes the code, and gets better at it every month. What it cannot do is decide what should be built, how it should be structured, or whether it will hold. This course teaches that part: the setup we use ourselves, and why every piece of it is there.

Software Engineering Principles for the AI-Native Builders

Join the waitlist See what's inside
Two ways to build with the same AI
Two ways to build with the same AI Exploring: a loop of prompt, look, adjust. Right for a prototype or an idea. Building to last: plan, design, build, test, deploy, maintain. Right for anything customers or money depend on. The course teaches the second and when the first is the right call. Exploring fast, throwaway, fine prompt look adjust Right for a prototype, a weekend, anidea you want to see. Building to last each stage leaves somethingthe next one reads Plan Design Build Test Deploy Maintain maintain feeds back into plan Right for anything customers, moneyor your name will depend on. Same tool, different discipline. The courseteaches the second, and when the first isthe right call.

You saw what AI could build. You built it. It works.

Can you trust everything the AI wrote? Do you need to understand all the code? How much of what's happening in the background do you actually need to know?

Fair questions, every one of them. You can pay someone to answer them once, or learn to answer them yourself, every time.

  • 01

    You're building a business on it

    This is not a side project. Customers, revenue or an investor are going to depend on what you build, and you want it to hold when they do.

  • 02

    You want to depend on yourself

    Not every question needs an engineer on the phone. You want to know which ones do, and handle the rest correctly yourself, the first time.

  • 03

    You want to hold your own with engineers

    When a developer, an agency or the AI itself proposes something, you want to judge it, ask the right question, and know when the answer is wrong.

The syntax was never the hard part

'Do I need to learn to code?' is the question we hear most. Our answer: not really. What separates a senior engineer from a beginner isn't syntax, it's judgment: what to check, how to structure a project so it doesn't collapse under its own weight, when to slow down and when to move fast. The industry is having exactly this argument right now, in public.

  1. 01

    Copilots, not autopilots

    The clearest statement of the distinction this course is built on: exploring with AI is the right way to find out what you want, and engineering with AI is what production needs. The difference is a set of practices, not a set of tools, and every one of them can be taught.

    Addy Osmani, 2025
  2. 02

    Judgment is the skill gap

    The real shift isn't that AI writes the code, it's that judgment (what to check, how to structure things, when to slow down) is now the whole difference. Almost nobody teaches that part on its own.

    Medium, 2026
  3. 03

    Spec first or build first

    'Spec-driven development' versus 'vibe coding' is a live argument among software teams, not a framing invented here. Knowing which fits which moment, and why the honest answer is usually both, is the first module.

    InfoWorld, 2026

What you'll learn

Six areas, not six weeks. You build your own app as you go, in order or by jumping to whatever it needs today. Each area is a piece of the setup we build with every day, and the reason it is there.

What holds an app up: the part the course is about
A building above ground and its foundations below Above ground: what customers see, what your team uses, the features. Below ground, the piles that hold it up: the data model, access rules, structure, tests, backups and recovery. ground what customers see what your team uses the features what people use 1 2 3 4 5 what holds it up 1 data model 2 access rules 3 structure 4 tests 5 backups

As deep as you need, no deeper

There is a lot of material, and some of it gets technical. Every area is layered so you can stop at the level your app needs today and come back for the rest.

  1. Level 1

    Enough to decide

    What it is, why it matters, what goes wrong without it. An hour.

  2. Level 2

    Enough to do it

    The setup, step by step, applied to your own app. An evening or two.

  3. Level 3

    The full picture

    What is happening underneath, for when you want to understand rather than follow.

Two weeks of evenings covers the level you need on the areas that matter to your app. The rest is there when you want it.

  1. Area 01

    Two ways to build, and when each fits

    Exploring fast to see an idea, and building to last once the idea is real. Both are legitimate. The expensive mistakes come from doing one while thinking you are doing the other.

    Why it matters   Same tool, different discipline. Knowing which mode you are in is the first skill.

  2. Area 02

    Say what you want before you ask for it

    Intent and specification, written in plain language before any code: what the software is for, who uses it, what must always be true. The AI builds exactly what it is asked. This is how you ask well.

    Why it matters   The expensive mistakes are in what the AI was never asked.

  3. Area 03

    Foundations before features

    How the data is shaped, who may see and change what, and where the boundaries between parts sit. The decisions that are cheap in week one and expensive in month six.

    Why it matters   A feature is an addition. A foundation change is a renovation.

  4. Area 04

    A professional's setup, explained

    Version control so any change is a five-second undo. Development, staging and production kept apart. Secrets out of the code. And the AI's own workspace: project rules it follows, checks that run on every change, prompts built to make it verify rather than agree.

    Why it matters   The setup is what catches the confident wrong answer before anyone depends on it.

  5. Area 05

    Test, ship, and keep it alive

    Tests on the flows that matter, a release you can undo, backups that have actually been restored, alerts when something breaks, and a record of what went wrong that feeds the next plan.

    Why it matters   Production is where the software meets reality. This is how it survives the meeting.

  6. Area 06

    Audit your own work

    The checklist behind a paid audit, taught as a skill: what to check, how to tell 'it ran' from 'it is right', and how to write down what you looked at.

    Why it matters   'No finding' and 'didn't look' are never the same thing.

    The same list we run as a paid audit, if you would rather have it done for you.

What this is not

A coding course. You will read code and understand what it does, but you will not be asked to write it. The AI does that, and gets better at it every month. You will learn the decisions around the code, which is where the difference between a demo and a product is made.

That judgment usually takes years of expensive mistakes to build. We teach it directly, for builders already shipping real products with AI. This won't make you a senior engineer. It closes most of the gap, and teaches you exactly what's left, so you know when it's worth calling someone for the rest.

The setup is the one we build with

We build with AI every day, and the process this course teaches is the one we ran on BizBuy and Epic Trading, both live products in the work on the home page, taking each from prototype to production. This isn't theory assembled for a cohort. It's what those systems needed, taught directly instead of delivered as a report.

Be first in when it opens

The course is being built now. Leave your email and you will know the moment it opens.

  • First access when it opens
  • One email, when it opens
  • Nothing else in between

Joining puts you on our email list: we write when the course opens, and occasionally about what audits turn up. A few times a year at most, and one click stops it for good.

We keep your name and email to write to you, and delete them after 24 months. Nobody else sees them, ever. Privacy notice.

Free to join. Leave whenever you like.

Prefer email? Write to us and we'll add you to the list ourselves. [email protected]