Team training:
Build with AI
Book a callWhich format suits your team?
Same workshop, same content. You choose the split.
What your team takes away
One setup for everyone. Project rules in the repo. A review process for AI code.
And the plan for the first 30 days.
In most teams the experience is spread very unevenly: some have been experimenting with agents for months, others have never used one. By the end of the training everyone works with the same toolchain. You use it to build features, find bugs, write tests and modernise legacy code. We prefer Claude Code and compare Codex, Antigravity and Cursor against it.
For your whole team
Backend, frontend, legacy: you bring your stack, we adapt the sample project.
Your own agent from day 1
From the very first hour, your own agent works on a complete sample project.
No AI experience needed
Programming experience is enough. We work out everything else together.
Live online, worldwide
Live and interactive with your team, no travel needed: hands-on exercises, open discussion and time for your questions.
Four days of 5 hours each
For example 9am to 3pm including breaks. Between the sessions you put what you learned to work in your day-to-day business.
Learning in small groups
We recommend up to 12 people per course, which leaves time for your questions. For larger teams we suggest splitting up, for example into 2 × 10; one joint course works too.
4 days of AI training
Online
From agent to team workflow.
Everything for productive work with AI agents, from Claude Code to your own MCP server.
As an online training: 4 × 5 hours.
Topics & content
BLOCK01- How LLMs tick: tokens, context windows, limits
- Where agents fail and how to spot it early
- Claude Code: setup and first tasks Hands-on
- Project memory with CLAUDE.md
- Delegating tasks the right way
- Letting agents work on the sample project
- Reading and evaluating the results
Understand & get going
Models, setup, first tasks
First we look at how language models work and where agents reach their limits. Then you set up Claude Code and let an agent work on the sample project. By the end of day 1, everyone has done this themselves and understands why the agent does what it does.
- Context engineering: the right knowledge at the right time
- Spec-first: specification before code
- Reviewing plans before the agent gets going
- Guardrails for agents
- Tool comparison: Codex, Antigravity, Cursor Live
- Reviewing AI code: typical classes of bugs
- Tests as a safety net
BLOCK02Stay in control
Context, plans, review
Agents deliver usable code when they know your project. You'll learn to prepare project rules, specifications and plans so they can find their way around. And you'll learn to check AI-generated code systematically.
BLOCK03- Understanding MCP: the power socket for AI systems
- Using MCP servers: browsers, databases, internal APIs
- Building your own MCP server (TypeScript SDK) Hands-on
- Skills: encapsulating recurring tasks
- Subagents and parallelisation
- Multi-agent workflows
Extend & connect
Plugging agents into your world
The Model Context Protocol (MCP) connects agents to your systems: databases, browsers, internal APIs. In the course you'll build your own MCP server and learn to encapsulate recurring tasks in skills and let several agents work in parallel.
- Agents in CI/CD: GitHub Actions
- Automated code reviews
- Headless: tasks that run overnight
- Legacy refactoring with agents Hands-on
- Team workflows and knowledge sharing
- Understanding and controlling costs
- Roadmap: your first 30 days afterwards
BLOCK04Automate & anchor in the team
CI/CD, reviews, team workflows
In the last block agents work unsupervised: as automated code review, in the CI pipeline and overnight for refactorings. In the hands-on case of legacy modernisation you'll bring it all together and take home a concrete roadmap for how your team anchors these workflows in its first 30 days.
All participants receive a certificate plus all the setups, configs and cheat sheets from the course to take home.
Your trainer
Johannes runs every training himself, with the knowledge from his ongoing projects.All workshops since 2016 in the Hall of Fame →
Johannes Hoppe
is a Google Developer Expert (GDE) and works as a freelance software architect, consultant and trainer with a focus on agentic coding, Node.js, .NET and Angular. He is co-author of the standard German-language book on Angular. Alongside the trainings he works as a lead frontend developer in client projects and applies what he teaches there.
Johannes lives in Leimen near Heidelberg.
Bluesky: @johanneshoppe.de
LinkedIn: @johanneshoppe
Frequently asked questions
◢ What does the training cost?
Online €6,600 net (4 × 5 hours), on site €8,250 net (3 × 7 hours) including travel, expenses and accommodation. It is a fixed-price offer for your team; the actual number of participants does not change it. For comparison: with twelve participants that is €550 or €688 per person.
◢ Who is this training for?
Whole development teams, backend as well as frontend. Programming experience is enough, no AI experience needed. We align the sample project with your stack; without a preference we use TypeScript. Everyone needs their own laptop; we send the installation guide well in advance.
◢ Which tools do you cover?
We prefer Claude Code. Alongside it we compare Codex, Antigravity and Cursor with their strengths and weaknesses. Trial accounts are enough for the exercises; what continuous use costs and how to keep those costs under control is part of the course itself.
◢ How does the hands-on part work?
From day 1 an agent works for you on a sample project. We explain each topic briefly and put it into practice right away; selected exercises you work on in small groups. Your company code stays with you. On request we look at it beforehand and set the focus accordingly. For AI features in your own product there is Build for AI.
◢ Are there learning materials for the training?
For every topic block we prepare a handout. Our detailed, up-to-date slide decks let you follow every topic comfortably. You get all the slides after the training. You'll also take home all the setups, configurations and cheat sheets from the training and can reuse them in your own project.
◢ When and how long?
We agree on the dates with your team. Online it is four days of 5 hours each (e.g. 9:00 to 15:00), on site three days of 7 hours each, breaks included. Between the online sessions you try out what you learned in your daily work.
◢ In which language is the workshop held?
In German or English, as you prefer; the course materials are available in both languages.
◢ What about data protection and confidentiality?
You do not need any company data for the exercises: you practise on the sample project, with trial accounts for the AI providers. Your code and your data stay with you; for the code review beforehand you send us only what you want to show, we treat it confidentially and delete it after the training. If only certain providers or local models are allowed at your company, we align the exercises accordingly. From the participants we only need names for the certificates. If your compliance requires a non-disclosure agreement or a data processing agreement, just let us know.
Still have questions?
+49 151 681 222 66 (Johannes Hoppe)
Standard rate to German mobile networksWe'll get back to you within 24 hours.Book a call
Tell us briefly where your team stands. We’ll get back to you with suggested times.
- Call30 minutes, free of charge. We find out where your team stands and where to put the focus.
- Quote within 24 hoursOne fixed price for your team.
- Dates from your calendarWe suggest dates that suit your team, online worldwide or on site in Germany, Austria and Switzerland.