Apprentice wanted: security and SaaS setup, learned by doing it

Illustrated infographic summarizing: Apprentice wanted: security and SaaS setup, learned by doing it

By Greg Nowak. Last updated 2026-10-10.

I learned server operations by breaking my own servers and fixing them again. It is still the way I believe in most. So I am not looking for an intern to sit with PowerPoint. I am looking for someone who will stand next to me while I do it, then do it themselves the week after.

What you will get your hands on

I run a server fleet with around 50 domains behind Cloudflare, Drupal and Laravel sites, a password manager, mail routing, backups and a handful of AI agents working for customers. All of it is under attack around the clock. Most of it bounces off because someone set it up properly. That someone can be you.

Week one: you get access, and we check one of my own domains together. Headers, TLS, versions, logins, backups. Week two you do it alone on the next domain, and I read your report. Then come the customers' sites. Later you set up tools for a customer: a Trello board, Google Workspace or Microsoft 365, two-factor access, a password manager. And one day a real problem lands in the middle of it all, and you watch how it gets solved.

You get your own account with two-factor, and only access to the system the task is about. Customer sites come after you have run the checklist on my own domains, and I look at anything you change on a customer system before it goes in. Customer data stays where it is. When we stop, I close your accounts.

What I expect

That you can find your way around a terminal. That you know what DNS does, or will find out the same day. That you write down what you did, briefly enough that I will read it. And that you say it out loud when you have made a mistake. Mistakes are expected. Hidden mistakes are the only thing that is not okay.

You do not need certifications. I would rather see a server you set up or a script you wrote than a course certificate.

The terms

We start with one day a week and can go up to two. Remote, with a fixed weekly call. It can be a paid student job, or it can be an internship through your studies if your programme requires one. We decide that together once we have talked. Three months to begin with, then we see.

You can show what you did here afterwards. First we take out customer names, domains, IP addresses and anything an attacker could use, and I read it before it goes anywhere.

Write to me

Use the contact form with the subject [grn-praktik]. The subject is filled in when you come from this link. Send me a link to the most broken thing you got working again, and two lines on what went wrong. I answer myself.

Latest articles

An internal AI can keep quoting a deleted policy. Check document updates, indexed passages and cached answers to find where outdated guidance survives.

Codex drew all ten pages of my Ninjai comic from the script Claude wrote in the first test. What made that handover work, and what it cost.

ChatGPT delivered ten painted comic pages; Claude drew one page in code. A practical look at AI tool selection, creative briefs and approval decisions.

Get missing supporting documents under control with a clearly scoped AI workflow that sorts replies, coordinates reminders and gives the bookkeeper a clear follow-up list.

Know what dmesg and journalctl can tell you after a reboot, how to find previous-boot logs, and how to check that retention covers your investigation window.

Check how your chatbot identifies itself as AI, where requests for human help actually go, and what customers encounter when support is unavailable.

Start with one website change on board.nowa.dk. Write a clear task, answer questions as they arise and review the result before giving an AI agent more work.

Keep AI search visibility, training preferences and assistant access separate. Use a practical policy matrix to check Cloudflare settings and test key public pages.

Let AI draft Black Friday campaigns using approved product data. Check prices and recipients, and secure approval before sending the campaign from Mailchimp.

Make AI meeting notes useful after the call: confirm decisions, agree owners and deadlines, check existing tasks, and verify that work reaches the right tracker.