I spent five years in DevOps and SRE at a large bank, working on incidents, automation, observability, Kubernetes, and CI/CD. After a layoff, I took a better-paying contract that turned out to be mostly ticket handling, runbooks, and escalations.
How can I keep building SRE skills and explain this role when applying for engineering jobs? I am also considering an AWS generative AI certification for AIOps. Is that a useful direction?
Look for toil you can reduce within the access you actually have. Even a small script or reporting dashboard can be useful if it makes recurring work easier to spot or handle. Keep track of what changed and any impact you can reasonably measure.
That gives you something concrete to discuss as SRE practice, even if you are not allowed to make application-level fixes. Just be clear about the scope rather than presenting a support dashboard as ownership of the whole system.
I would not hide the support work. Handling incidents, communicating with users, following escalation paths, and explaining technical issues clearly are all relevant skills. They are different from owning infrastructure, but they are not worthless.
On the resume and in interviews, distinguish the current contract from your earlier engineering work. Describe what you actually did in each role, then explain that you are looking to return to a position with more ownership and building.
For practice outside work, pick one small service and treat it like a system you have to operate. Define a reliability target, add metrics and alerts, run a load test, and write down what you would change when it misses the target. You could also introduce a controlled failure and practice recovering from it.
The useful part is the reasoning, not how many tools you can squeeze in. Keep notes on assumptions, tradeoffs, and what your tests did or did not demonstrate. A home lab is not the same as operating a trading system, so frame it accordingly.
I would treat the generative AI certification as an interest to explore, not a guaranteed route back into senior SRE. Anomaly detection and automated remediation are relevant ideas, but the certificate alone does not establish that you can design or operate a reliable production system.
If study time is limited, compare the certification material with the requirements in roles you want. You may get more direct practice from reliability fundamentals and a concrete project. If the AI material genuinely interests you, there is room for both.
Answers are reviewed before publication.
under-a-minute:
I have put together reporting dashboards, but our team has limited application access and not enough context to investigate or fix bigger issues.
plaid without a plan:
Then document what the dashboards made visible and where the team had to escalate. You can still talk about identifying recurring toil without claiming you fixed its underlying cause.