Learn to check AI-generated code and deliver changes you can stand behind.
£495 including VAT per place.Four live 40-minute sessions.
Each of the four sessions teaches one operating rule, applied to a realistic engineering task in the same session.
For engineers with two months to two decades of experience. You should be comfortable reading, changing and testing code. Newer engineers get a method to follow. Experienced engineers examine decisions they may otherwise make by instinct. No particular model or editor is required.
Work mostly in documents and data? See the business course.
The examples below come from the current course repositories. They describe the learner task without revealing the facilitator checks.
Until you have checked it independently, generated work stays untrusted.
Ask what the request means before you start implementing it.
A coding tool gets only as much work as you can still properly review.
Change course when the evidence shows that the current approach is wrong.
Four live 40-minute sessions, weekly over four weeks, each built around prepared repository work with live tutor review.
Each repository includes a task, existing code and visible tests. You keep the final change and written work you produce.
You can take all four sessions in Python. Some sessions also have TypeScript, C# or Rust exercises, and Lee will confirm which set your cohort uses before you book.
Lee explains the rule and the engineering risk behind it.
You inspect the task, use your coding assistant and run your own checks.
You explain what you accepted, changed or rejected and show the evidence.
Compare your approach with the rest of the group and discuss where you'd use it at work.
Lee looks at your finished change and the checks behind it. He also reviews any extra changes you allowed in, and what you did when a check challenged your approach.
Model choice, prompt style and token use are not scored. A single session gives limited evidence and is not presented as a certification.
Representative tutor feedback, written from the assessment rubric
The tests you added catch the reported case, but the supplied patch also changes the retry path and nothing exercises it. Show the check that would fail if that behaviour regressed, or take the retry change out of the change you keep.
The first engineering cohort ran in July and August 2026. These two comments came through the course feedback form with permission to publish, and both appear in full.
The AI-Assisted Engineering course is exceptionally well structured and refreshingly practical. Its focus on defining work clearly, independently verifying AI-generated code, controlling scope, and adapting when evidence challenges an approach addresses the realities of using AI in software development. The hands-on repository exercises, live tutor feedback, and support for multiple programming languages make the learning immediately applicable. It’s a thoughtful, credible course that promotes responsible engineering judgment, not simply faster code generation.
Camila CotaEngineering course, September 2026Before the course, I mainly focused on solving the technical problem. Now I ask better questions, think more like a product owner, and look for evidence first. I learned to check test cases, reproduce issues, keep logs, and make small, meaningful commits. This has made my development process more structured and reliable.
The practical exercises helped me most because they showed me how to apply these habits to real problems while using AI responsibly and validating its output.
I especially appreciated how supportive you were throughout the course. You always took the time to answer my questions and guide me, which made the learning experience much more valuable.
One thing I would improve is adding a collaborative project where participants work together on a larger task. I think that would make an already great course even better.
Sujal NeupaneEngineering course, September 2026Small, well-defined tasks with a quick check are where coding assistants can be most useful. Long, ambiguous repository work can hand that gain back through review, rework and regression cleanup.
A controlled experiment found developers completed one defined JavaScript HTTP-server task 55.8% faster with GitHub Copilot.
METR's early-2025 randomised study found experienced open-source developers took 19% longer with AI tools on real issues in repositories they knew, even though they believed they had been faster.
Lee calls the lost time review churn and expansion: plausible changes grow, engineers find regressions and the time saved generating code is spent establishing what is safe to keep.
Read the controlled Copilot study, METR's early-2025 study and METR's 2026 update. The later update says newer tools probably improve results, but selection effects prevented a reliable current estimate.
The course is live and unrecorded. You control what you share with Lee and the group.
Get the dates when booking opens.
Get the engineering course datesRead the waiting-list terms or booking terms.
