Live cohort • Live online • Next cohort
Walk out with a production-grade agentic engineering setup running in your own codebase.
Four live sessions. Your real codebase. You walk out with agents running in production — not a side project you'll finish someday.
What you walk out with
A working artifact, not a slide deck.
- →
A CLAUDE.md and rules-file structure tuned to your codebase
- →
At least two custom skills you build during the cohort
- →
Sub-agent definitions, MCP server configs, and IDE integrations
- →
Automated checks that catch AI output problems before they reach production
- →
An evaluation framework you keep running after the cohort — not just on day one
- →
Validated workflows you apply in the next sprint — not someday
What this is not
Architectural patterns, not framework-of-the-month.
- Not a LangChain bootcamp. Not a CrewAI tutorial. Not "agentic ChatGPT" rebranded.
- We focus on architectural patterns — context engineering, sub-agents, MCP, evaluation, governance — that outlast any specific framework.
- Tool choice is a downstream decision. We make it last, not first.
This is for you if
Product developers, full-stack engineers, technical PMs who code.
- You use AI tools but the value is uneven — sometimes magic, sometimes worse than typing it yourself
- Your pull requests look "AI-assisted" but quality is inconsistent
- You want agents in every development phase — refinement, planning, implementation, review — not just code completion
- You want this to be team practice, not just your personal experiment
Curriculum
4 sessions × 2h over 2 weeks of practice — built around your real product.
- Module 01
Context engineering: the thing that actually matters
LLMs vs. agents. Why most AI-coding failures are context failures. CLAUDE.md, rules files, project-wide configs. Context window mechanics. How to give agents the information they actually need.
- Module 02
Agent architecture, not agent hype
MCP servers connecting agents to your real stack. Sub-agents for parallel execution. Skills as reusable team workflows. IDE integration. Patterns that outlast your next tool migration.
- Module 03
From demo to production: evaluation that keeps running
Apply agents to features and bugs from your own backlog. Build a measurement system that catches AI output problems before they reach production — and keeps running after the cohort ends.
- Module 04
Agents across every development phase
Refinement, planning, implementation, review. Hooks and automated quality checks as production guardrails. A governance model: which agent decisions need human review, which don't.
Instructor
Practitioner. Not a training brand.
Twenty years shipping product — the last ten as a scrum master and agile coach inside teams, not as a consultant billing days. I run AI in production on my own products. Cohorts stay at max 8 so everyone gets time on their real work, not a made-up exercise.
Format & what's included
Practical details.
- Format
- 4 sessions × 2h over 2 weeks, live online via Zoom
- Cohort size
- Max 12 participants
- Price
- From €1,200 excl. VAT
- Next cohort
- — confirmed 4 weeks before launch
- Included
- Pre-work + post-work template · session materials in your repo or workspace · written summary of decisions and patterns
Reserve a seat
Want this for the cohort?
Be one of the first 6 people through this cohort. Reserve your seat and we'll confirm dates once 6+ people join — usually 3–4 weeks before launch.
No commitment required. We confirm cohorts only when 6 or more seats are reserved.
If the cohort doesn't fill, you're automatically released — no charge, no hassle.