Training
Hands-on courses on running your own infrastructure: Kubernetes with Gardener, Hetzner Cloud, Forgejo and CI runners, and shipping a small SaaS. Every participant works in their own browser workspace with real servers behind it. Nothing to install.
Hetzner Cloud crash course Pilot
CLI, Terraform, Python · 3 × 2 h live + 3 h homework
Operate Hetzner Cloud the three ways teams actually do it: the hcloud CLI, Terraform, and Python scripts — against a live mockup of the Hetzner Cloud API with real virtual machines.
On-demand CI runners Pilot
A runner per job, and a bill that stops · 3 × 2 h live + 3 h homework, or on demand
Run Forgejo CI runners that exist only while a job needs them: a server per waiting job, a single-use registration per attempt, and deletion before the next paid period starts.
Running Forgejo in production Pilot
Move off GitHub without losing your CI · 9 self-paced workshops
Migrate a repository with its issues and CI workflow into a Forgejo of your own, then run it: runners, backup and upgrades, SSO, scaling and cost, and the real migration for a whole organisation.
Running Kubernetes with Gardener Planned
Day two for small teams · 3 weeks, about 20 h live
Operate the clusters Gardener hands you: restore backups, watch upgrades, wire an app with ingress, TLS, storage and alerting, run a game day, and write your own runbook.
Ship a SaaS in 3 Weeks Planned
Build and launch your own small AI SaaS · 3 weeks, 7.5 h live + homework
Build your own Django SaaS with a coding agent: a core feature, one AI feature, payments with Polar, and a production deployment on Hetzner — live at your domain after three weeks.
Building an AI Software Factory In preparation
Eight labs with Claude Code · 8 live sessions over 4 weeks
Turn a repository only its author can build into a small software factory: reproducible build, work packets, sensors and hooks, a sandbox without the key, holdout verification, an orchestrator, budgets and a review policy.
AI agents in operations In preparation
Let an agent diagnose and act, within guardrails · 2 weeks, about 12 h live
Let AI agents help operate self-hosted apps: what an agent may read and what that leaks, typed actions with approvals, restore drills and updates with rollback run by an agent, and measuring an agent before trusting it.
How it works#
- Browser workspace. Each session opens an editor and terminals in the browser (Educates), with the tools installed and the lab environment already running. Nothing touches your laptop.
- Pilot phase. The first sessions run on my own servers, so seats are limited. During the pilot they are free.
- Later. Sessions will be paid; the price is on the page before you book.
- Pre-registration is an email to me. I write back when the first sessions are scheduled.