Why I Completely Rewrote My DevOps Course for 2026
For the last five years, I’ve taught a DevOps course at Harbour.Space University in Barcelona. The curriculum evolved naturally, but the core remained steady: version control, CI/CD, infrastructure as code, observability.
But in early 2026, I hit a breaking point. I realized I don’t write code by hand anymore.
My daily workflow as an SRE Manager has completely changed. AI agents are now smart enough to handle complex tasks—investigating bugs, writing Terraform, and creating sophisticated pull requests. Teaching students to write boilerplate YAML or shell scripts from scratch suddenly felt like teaching them to use a slide rule.
So, I rewrote the course. The 2026 curriculum is now AI-Assisted DevOps.
The Fundamentals Haven’t Changed (But How We Apply Them Has)
A common misconception is that AI replaces the need to understand how systems work. It doesn’t.
The fundamentals are largely the same. You still have to run builds. You still have to deploy to production securely. You still need to understand networking and Linux. What has changed dramatically is that AI can now assist with all of these steps.
The new bottleneck isn’t typing speed or remembering syntax; it’s judgment.
To use AI effectively, you have to ask the right questions. You have to prompt it properly. Most importantly, you must be able to judge whether the output is actually good enough to run in production. You can’t evaluate an AI-generated deployment pipeline if you don’t understand the fundamentals of how a pipeline works.
The Mindset Shift
When my students finish the 3-week course this June, I want them to walk away with a specific mindset shift: AI is your primary learning and execution tool.
I am equipping them with practical examples of how to build and deploy pipelines using AI agents, while maintaining complete understanding and control over what is going on under the hood.
The goal isn’t to make them dependent on AI, but to teach them how to direct it. The future of DevOps isn’t about writing the code; it’s about architecting the system and letting the agent handle the implementation details.