How the protocol works
- Hosts, clients and servers, and which one you are writing
- Tools, resources and prompts, and when each is the right fit
- stdio and HTTP, and what changes between them
- Initialisation, capability negotiation and shutdown
Training course
Advanced
Tailored to your team
MCP Server Training
Build the layer that lets a coding agent reach your own systems, and keep control of what it can do with them.
- Level
- Advanced
- Length
- 1 to 2 days
- Delivery
- On site, Remote, Hybrid
- Group size
- 5 to 20 participants
Booked privately for your team. Dates and length are agreed when you book.
What your team will be able to do
Every module ends with something participants can repeat on Monday morning, in your codebase and against your conventions.
Build a server that runs
Stand up an MCP server from an empty directory and have an agent calling it before lunch.
Expose your own systems
Put an internal API, a database or a document store behind a tool an agent is allowed to call.
Design tools agents can use
Name, describe and shape a tool so the model picks it correctly instead of guessing at it.
Pick the right transport
Decide between stdio and HTTP for a given server, and know what each one costs you operationally.
Keep credentials contained
Hold secrets outside the agent's reach and scope every tool to the least it needs.
Get it onto the team's machines
Version, package and roll out a server so colleagues run the same one you built.
Curriculum
Built around hands-on work rather than slides. The running order and the depth of each module are adjusted to the group before delivery.
75% hands-on
Your first server
- From an empty directory to a server the client can see
- Declaring a tool and returning something useful from it
- Serving read-only context the agent can pull on demand
- Driving the server by hand before an agent ever touches it
Designing tools an agent will use well
- Names and descriptions the model reads as instructions
- Input schemas that make a wrong call impossible
- Failing in a way the agent can recover from
- Trimming responses that quietly eat the context window
Running one for a team
- Authenticating the server against the systems behind it
- Where credentials live when the agent must not see them
- Getting the same server onto every machine
- Changing a tool without breaking the agents already using it
Built around your team
No two runs of this course are the same
The curriculum above is a starting point, not a syllabus. Before anything is booked I find out where your team actually stands, then rebuild the course around that. A team already running agents daily spends its two days somewhere completely different from one that installed the tool last week.
-
Step 1
Intake call
A call before anything is agreed: what your team already uses, what has gone wrong so far, and what you need them doing differently afterwards.
-
Step 2
Stack review
Language, framework, review process, CI. Exercises are rebuilt against your repository so nothing in the room is hypothetical.
-
Step 3
Curriculum reshaped
Modules get dropped, deepened or added. You see the adjusted outline and sign it off before the date is fixed.
-
Step 4
Delivery and follow-up
We run the course, then check in once the team has applied it to real work and the awkward questions have surfaced.
What gets adjusted
- Depth of each module
- Your stack and repository
- Group size and seniority mix
- Length and how days are split
- English or German
- On site, remote or hybrid
Is this course right for your team?
Who it is for
What participants need beforehand
- Working knowledge of Python or TypeScript. The exercises run in whichever your team uses.
- Day-to-day use of a coding agent already behind you. This is not an introduction to agents.
- A laptop with an agent client installed and permission to install packages on it.
How the training is delivered
Same material either way. The format changes how much of it lands in the first week.
On site
I come to your office and work with the team in the room. Usually the most effective option for a first rollout, because the sideways questions get answered as they come up.
Remote
Delivered live over video in focused blocks rather than one long day, with shared screens and hands-on exercises on your own work.
Hybrid
A day on site to get everyone moving, followed by shorter remote sessions once the team is applying it to real work and hitting real friction.
Optional add-ons
Certificate of participation
There is no public schedule. Every course is booked privately for a single team, and we agree the dates together.
Your trainer
Luka Breitig
Software engineer and AI coding trainer
I train engineering teams to work with AI coding tools, and I use the same tools to ship production software every week. The material comes out of building things, not out of a slide deck.
Before going freelance I spent six years as technical co-founder of a marketing agency, responsible for everything technical from automation infrastructure to coordinating a freelancer network. That is why the courses spend as much time on review, guardrails and rollout as they do on prompting.
- Ships production software every week
- Delivers in English and German
- On site across Europe, or remote
- Speaker at Nomad Summit, Chiang Mai
Questions about this course
The things teams ask before booking. If yours is not here, ask me directly.
No, and most groups have not. What the course assumes is that you already use a coding agent and have run into something it cannot reach. We start at the protocol and have a server running within the first two hours.
Python or TypeScript, whichever your team writes. The protocol is the same either way, so a mixed room is fine and usually instructive.
That is the intended shape of the course. Send me one or two systems you would want an agent to reach beforehand, and the exercises are built against those instead of a sample API.
No. Everything runs locally on the participants' machines. Hosting and distribution are covered in the last module so you leave knowing the options, but nothing has to be provisioned beforehand.
Very little. The Claude Code course covers using an agent well, including connecting a server someone else has written. This one is about writing the server, and it assumes that other side is already familiar.
Five to twenty. Debugging a protocol handshake is a one-to-one conversation when it goes wrong, and past twenty there is not enough of me to go round.
Still not sure whether this is the right fit for your team?
Book this training for your team
Tell me the size of the group, the codebase they work in, and roughly when you want to run it. I will come back with a proposal.
Delivered in English or German, on site across Europe or remotely.
Trusted by teams & training providers
I deliver AI training for some of Europe's leading organisations and L&D platforms.
Cegos Integrata
Commissioned trainer via leading European L&D provider
NobleProg
AI coding workshops via one of the world's largest IT training marketplaces
Innomotics
In-house AI training for engineering teams
Bots & People
Commissioned trainer for developer workshops
Code First Girls
Course creator — 3-month AI & agentic tools curriculum
Nomad Summit
Conference speaker on AI tools, agentic development, and developer workflows (2025)
Vertiv
In-house AI-assisted coding training for engineering teams