AI for development teams
We teach your engineers to use AI coding tools well. Your team learns a practical method on your own codebase, and you measure the result.
Talk about your project§ 01What you get
The result, in plain terms.
- A training program that uses your codebase, your stack and your tools.
- Shared project instructions and conventions that every developer can use.
- Clear rules for code review, security and data.
- Coaching on real tickets, so the method stays in use after the training.
- Measurements of delivery speed and quality before and after.
§ 02The curriculum
A practical course, taught on your own code.
Module 01
Foundations
How the models work, what they need in context, and where they fail.
Module 02
Project instructions
Rules files, conventions and shared commands that give each tool the right context.
Module 03
Task design
Break work into steps that an agent can do and that a person can check.
Module 04
Review and testing
Tests first, a strict review of generated code, and automated security checks.
Module 05
Agent workflows
Multi-step tasks, agents that work in parallel, and AI in the CI pipeline.
Module 06
Team practice
Policy, data safety, cost control and the measures that show progress.
§ 03The route
Stage 1
Survey
We look at your codebase, your tools and your current AI use. We ask developers what works and what does not.
Stage 2
Plan
We adapt the curriculum to your stack and your rules. We agree on what to measure.
Stage 3
Teach
We run hands-on workshops on your own code. Each developer uses the method on a real ticket.
Stage 4
Embed
We pair with developers on real work. We set up the shared instructions and review rules that keep the method in use.
§ 04Ways to start
| Engagement | Scope | Typical length |
|---|---|---|
| Team assessment A review of current AI use, tools, policies and risks, with a plan. | A review of current AI use, tools, policies and risks, with a plan. | 1 week |
| Workshop series Hands-on sessions on your codebase, from daily use to agent workflows. | Hands-on sessions on your codebase, from daily use to agent workflows. | 2 to 4 weeks |
| Embedded coaching Pairing, review support and updates as the tools change. | Pairing, review support and updates as the tools change. | Monthly |
§ 05Questions
Which tools do you teach?+
We teach the method first and the tool second. Most teams use Claude Code, Cursor or GitHub Copilot. We use the tools that your company approves.
Our developers already use AI. Why do they need training?+
Most developers use a small part of what the tools can do. The largest gains come from shared project instructions, good task design and a strict review process. We teach those skills.
Is AI-generated code safe to ship?+
Only after review. We teach a review process and set up automated checks, so generated code meets the same standard as all other code.
Next step