Getting started
Principles, tools, and setup — and the code if you want to skip ahead.
By the end of this module you will have a working bench: tools installed, repo cloned, and a clear picture of how the factory is put together.
Online course · ~6–8 hours
Build your own software factory.
Turn your personal AI gains into company-wide systems with a software factory. Owned and controlled by you while learning from every ticket.
You can’t buy one. That’s the good news.
Every factory is unique, so you can't just drop in someone else's factory.
Start from the principles and build your own factory by running your own practices inside learning loops.
With a software factory, your whole team can gain from every learning, while controlling the process and quality.What you’ll build
Tickets in. Verified code out.
In one day of guided tutorials, you'll have a running software factory with four learning loops.
- 01Triage
- 02Plan
- 03Execute
- 04Verify
- 05Learn
Then it runs again — with what it just learned.
Eight modules
Eight self-paced, code-along modules
Each module is a live build of a piece of the factory and includes source code and full instructions.
Build your context learning loop
Your factory learns your codebase and conventions.
By the end of this module you will have a loop that captures how your team works and feeds it back into the factory.
Build your skills learning loop
Your factory gets better with every ticket completed.
By the end of this module you will have a loop that turns finished work into reusable skills.
Automate Claude Code
Point a worker at real tickets, not a demo prompt.
By the end of this module you will have Claude Code running as a worker you can aim at live work.
Put Claude Code in a loop
Unattended work across a queue — with the factory in charge.
By the end of this module you will have Claude Code running across a queue of tickets, not a single task.
Verification, oversight, and cost management
The module a budget holder needs before they approve the spend.
By the end of this module you will have checks, human oversight, and cost controls a manager can sign off on.
Integrating with a ticketing system
GitLab or Jira as the source of truth.
By the end of this module you will have the factory reading and writing tickets in the system your team already uses.
Running a local LLM
A path off the hosted APIs when you need it.
By the end of this module you will have a working local-model path, and a clear view of when it is worth the trade-off.
Format
Pause and replicate.
This is a build, not a lecture series. ~6–8 hours, self-paced. Watch a step, pause, do the step.
Code-along, not a lecture
Screen plus face cam. Real-time build of each module. Voice-driven prompting you can pause and replicate.
Source, branch per module
Full git repo. Stuck? Skip ahead on the next branch. The finished factory is in the last one.
Notes and command reference
Written notes for every module, plus the commands you will actually type.
One hour with the presenters
A one-to-one online consultation with the people who run the factory. Bring your codebase, your constraints, your questions.
Questions
FAQs
Who is this for?
Established teams — not solo beginners. SaaS companies lifting delivery capacity without headcount. Agencies that want to productise AI-assisted delivery. Government and larger orgs that need team-level AI adoption with oversight and verification built in. You should be comfortable at a CLI and following a code-along.
What does it cost?
US$999 provides access to the course and the consultation hour for you and your team.
How current is the content, given how fast this moves?
It is taught from a live commercial software factory, not a recorded lab. As the tooling shifts, we update the repo and the notes. You are learning a method that has to survive change — that is the point of the loops.
What tools and costs do I need?
A Claude subscription or API key. Optional local LLM path if you want to keep inference on your machine. Module 6 covers running costs and how to keep them honest. The local-LLM bonus covers the off-API route.
Is this for individuals or teams?
Teams. A CTO or engineering lead can take it and bring the method back. One person can follow the code-along; the factory is meant to be run by a team, but works well for solo developers too.
When does it start?
The courses will be available in September 2026.
What happens in the consultation hour?
One hour, online, with the presenters. Bring a real constraint: a codebase, a ticket queue, a risk rule, a cost cap. We will look at how to stand the factory up in your shop — not walk through slides.
Software factory
Build your own.
Then run it.
Join the waitlist. US$999 founding pricing locked in. We’ll email you when it opens.
No payment now.