- Every developer has a different prompting routine
- Tickets and context are copied manually between tools
- Debugging means checking code, logs and data separately
- AI-generated code requires several correction rounds
- Useful workflows live in individual heads
- Tasks run sequentially even when they could run in parallel
AI workshops for engineering teams
Get your engineering team fluent in AI.
A hands-on Claude Code workshop built around your codebase, tools and delivery process. Your team learns how to turn tickets, bugs and reviews into connected workflows that produce tested, reviewable code.
12 years building software · Interim CTO experience · AI in production since 2022
Measurable team leverage
Increase your team's development speed.Ship production software 3–5× faster.
The gain does not come from typing faster. It comes from giving AI the right context, connecting the relevant systems and building workflows the whole team can use reliably.
the_ai_delivery_gap
More AI-generated code does not automatically mean more productivity.
Most teams adopt AI at the individual level before creating a shared system for planning, implementation, testing and review. Developers produce more code in less time, but the delivery process around that code has not changed.
Without the right context, workflows and guardrails, AI does not scale productivity. It scales bad code.
- Everyone develops their own way of working with AI.
- Agents make conflicting changes and developers block each other.
- More code reaches review, but testing and quality cannot keep up.
- Missing context creates inconsistent architecture and repeated mistakes.
- Important knowledge remains inside individual chats and workflows.
- Code becomes harder to understand, maintain and test.
- The team feels faster, while release throughput barely improves.
The goal is not to generate more code. It is to build a shared engineering system that helps the whole team ship better software faster.
curriculum
What your team will learn.
A practical curriculum for turning individual AI usage into a reliable engineering system the whole team can use.
- 01How to scope and decompose work before delegating it
- 02How to give agents the repository context they need
- 03How to establish shared rules and code standards
- 04How to turn recurring tasks into reusable skills
- 05How to build documentation and action MCPs
- 06How to connect tickets, documentation, logs, databases and infrastructure through MCP
- 07How to build ticket-to-PR and bug-to-fix workflows
- 08How to add testing, review and deployment gates
- 09How to create an AI-native delivery architecture that supports fast, confident deployments
- 10How to improve development speed through parallel agents and isolated worktrees
Claude should not guess how your team works. The repository should tell it.
deliverables
What your team walks away with.
Not another collection of slides. Concrete deliverables your team can use and improve in real development work.
Shared AI development setup
A common starting point built around your repository, stack and existing delivery process.
Repository rules for agents
A clear structure for making architecture, conventions and code standards available to Claude.
Reusable rules and skills
Patterns your team can reuse and improve for recurring engineering work.
Connected engineering systems blueprint
A practical direction for making the right tools, knowledge and infrastructure available within agent workflows.
Delivery workflow patterns
Shared patterns for ticket-to-PR, bug-to-verified-fix and review-ready changes.
Testing, review and deployment gates
A quality model for keeping AI-generated changes testable, reviewable and safe to ship.
Multi-agent and worktree workflow
A practical approach for running isolated workstreams in parallel without developers blocking each other.
First-week implementation plan
Concrete actions for applying the workshop to real tickets, bugs and pull requests immediately.
before_after
Same team. A completely different way of working.
The shift is not from writing code to not writing code. It is from isolated experiments to a shared engineering system.
- Shared rules and workflows live in the repository
- Claude understands the project and its standards
- Relevant systems can be used together
- Changes pass through testing and review gates
- Recurring workflows are reusable across the team
- Multiple agents can work safely in parallel
connected_workflows
The real shift is not better prompts. It is connected workflows.
The workshop shows what becomes possible when AI is part of the engineering process instead of a separate chat window.
Ticket-to-PR workflows
From ticket to a tested, review-ready change with fewer manual handoffs.
Bug-to-fix workflows
Debug across code, logs and data instead of checking every system separately.
Parallel development
Run multiple isolated agent workflows simultaneously without them interfering with each other.
what_we_offer
What's included in every workshop.
Team and workflow intake
We map your stack, repositories, current AI usage and the workflows worth focusing on.
Preparation around your stack
The curriculum and demos are adapted to your engineering environment and level of adoption.
Practical workshop
Concepts are shown through real engineering work, not a generic slide deck.
Reusable materials
Your team keeps the templates, reference material and concrete next steps.
formats
Pick the depth that fits where your team is today.
All prices are team rates. Up to eight participants are included.
Impulse
From €1,200
Teams that want a practical introduction and a shared picture of effective AI-assisted development.
- Team and workflow intake
- Custom examples
- Live demonstrations
- First-week action plan
Practical Workshop
From €2,800
Teams that want practical workflows, shared standards and a usable foundation.
- Preparation based on your stack
- Hands-on engineering workflows
- Reusable templates and material
- First-week adoption plan
Team Intensive
From €5,500
Teams that want to go deeper and start integrating the setup into their development process.
- Deeper implementation work
- More room for team-specific workflows
- Multi-agent and worktree patterns
- Follow-up call after three weeks
Up to 8 participants included · Each additional participant €250 · Travel expenses billed separately
Discovery call → team and workflow intake → preparation based on your stack → workshop → concrete actions for the first week
why_stefan
Taught by an engineer who ships with AI in production every day.
I am not a full-time trainer. I build software full time. Everything I teach comes from workflows I use myself to plan, build, review, debug and deploy production systems.
Engineering credibility
- 12 years building and shipping software
- Interim CTO experience at LawGames
- Built a multi-tenant white-label SaaS from the ground up
- Reduced CI/CD release cycles by 65%
- Migrated 32,000 accounts and 5M+ records from Bubble to Next.js
- AI in real production projects since 2022
honest_fit
A practical workshop for teams ready to change how they work.
Good fit
- Engineering and product teams with uneven AI adoption
- Teams that care about maintainable code
- Teams that want shared workflows instead of individual prompt habits
- Leaders who want more speed without removing engineering judgment
Probably not
- Teams looking for a generic AI keynote
- Anyone expecting AI to replace engineering judgment
- Teams unwilling to work on their actual processes
- Organizations looking for a one-hour magic demo
faq
Questions teams ask before booking.
How is this different from generic AI training?
Generic training usually explains tools and prompts in isolation. This workshop is built around engineering delivery: repository context, shared rules, reusable skills, connected systems, testing and review gates, and safe parallel development.
How is team-wide AI adoption different from individual tool usage?
Individual usage can make one developer faster. Team-wide adoption requires shared context, review standards, quality gates and workflows that prevent people and agents from blocking each other.
What results can our team expect?
The goal is shorter delivery cycles, fewer manual handoffs, more reliable AI-generated changes and a shared way of working the team can continue improving. The exact leverage depends on your codebase, current setup and level of adoption.
Can you help us start and also scale what already works?
Yes. The workshop can establish the first shared workflows or help an experienced team standardize and scale practices that already create value. The intake determines the right depth.
Is the workshop only for software engineers?
Engineering is the primary focus. Technical product managers, QA engineers, DevOps specialists and engineering leaders can also benefit when they are involved in planning, review, testing or delivery.
What is included in a workshop?
Every format includes an intake, preparation around your stack and workflows, live practical training, reusable material and concrete next steps. The deeper formats add more implementation time and follow-up.
Do we need Claude licenses?
For hands-on formats, participants need access to Claude Code. We clarify the exact setup during the intake so the workshop time is not spent on account administration.
Does this work with Cursor or GitHub Copilot?
Many of the principles transfer to other tools. The workshop itself uses Claude Code because it provides the strongest foundation for the repository-level and agentic workflows covered here.
Will you work with our actual codebase?
Yes, where access and confidentiality allow it. The preparation is based on your stack and workflows. We can also use a representative private demo repository when production code is not suitable.
What about source code, secrets and company data?
Data boundaries, permissions and secret handling are part of the workshop. We separate context from credentials and treat read and write access differently.
What if part of the team is skeptical?
That is normal and often useful. The workshop focuses on concrete engineering work, quality and control rather than hype or replacing developers.
What experience level is required?
The workshop can be adapted to mixed teams. A basic understanding of the codebase and normal Git workflows is enough for the practical formats.
How much preparation is required from our team?
Usually one intake call plus access to the relevant technical context. I handle the workshop preparation and confirm any participant setup in advance.
What team size works best?
The listed formats include up to eight participants. Larger groups are possible for an additional €250 per participant, depending on the format.
Is the workshop remote or onsite?
The half-day Impulse is remote. The Practical Workshop and Team Intensive are designed for onsite delivery. Other arrangements can be discussed during the discovery call.
What happens after the workshop?
Your team leaves with shared material, practical starting points and actions for the first week. The Team Intensive also includes a follow-up call after three weeks.
Ready to start?
Make AI part of how your team actually ships.
Tell me about your team, stack and current AI usage. I will tell you which format fits and where a workshop would have the most impact.