I'm a junior DevOps engineer working on Kubernetes clusters and Linux server configuration, using Ansible for automation and Terraform for provisioning. The workload is dominated by a constant stream of tickets, and the team expects everything to be completed as quickly as possible, so I'm being pushed to rely heavily on AI-generated solutions.
The problem is that I rarely have time to understand what I'm doing or explore the areas I actually want to learn: deeper Kubernetes and Linux knowledge, Go, Ansible, infrastructure design, and troubleshooting. I understand the basic purpose of these tools, but I want to build and break my own infrastructure so I can develop real practical skills.
I'm worried that after several years of experience, my résumé won't show meaningful growth if most of my work consists of asking AI for answers and rushing through tickets. How can I use AI productively without letting it replace my learning, and how can I create time to build deeper DevOps skills?
4 Answers
AI should be treated as an assistant, not an authority. Read and test everything it generates, especially Helm charts, Terraform, and automation that can affect production. If you cannot explain why a change works, you are not really learning from it. Use the output as a starting point, then verify it against documentation, your environment, and smaller experiments. The fundamentals are still essential because they let you recognize when AI has produced a dangerous or incorrect answer.
Using AI is not automatically a career problem. The problem is blindly copying answers without understanding them. Ask it to explain commands, compare alternatives, identify risks, or create tests and validation checks rather than just asking for a finished solution. For example, you could build checks for Ansible inventories, rendered templates, Kubernetes manifests, network settings, and Terraform plans. Over time, that turns AI into a tool for improving your engineering process instead of a substitute for engineering judgment.
An urgent workload does not automatically mean every task is truly urgent. If everything is marked as ASAP, ask your seniors to help prioritize and explain the trade-offs. Look for recurring tickets and turn them into documented procedures, tests, or safe automation. That gives you learning opportunities during normal work instead of only trying to study after an exhausting day. Also keep notes on the incidents you investigate, the decisions you make, and the improvements you deliver; those are useful résumé and interview stories.
Set up a homelab and recreate the kind of environment you manage at work. Run Kubernetes locally or on inexpensive machines, manage it with Terraform and Ansible, and deliberately break things so you can practice recovery. You can also use open-source alternatives for tools that are unavailable outside work. Even a small project that you document in a repository can demonstrate much more than simply listing technologies on your résumé.
That is also how I learned while working a busy support role. I handled tickets during the day, then built a smaller version of the same stack in my own environment. It helped me connect the automated answers to the underlying concepts.
AI can produce convincing but incorrect results even for simple tasks. I once asked it to automate spreadsheet formatting and it changed unrelated cells, so verification still mattered. The same kind of mistake in infrastructure can become technical debt or an outage.