I spent about seven years working in internal IT and systems administration, managing services such as Google Workspace, Okta, and mobile device management. Three months ago, I moved into a platform engineering role at the same company.
I am the only platform engineer besides my manager, who originally built most of the current setup. Our environment is fairly decentralized: software engineers own and operate the infrastructure for their applications, while my manager and I mainly provide higher-level support. This gives me time to learn, but it also means there are not many obvious projects for me to work on because most of the platform is already established.
The biggest challenge is that I feel I lack important fundamentals. I can use AI tools to troubleshoot and complete tasks, but without that help I often struggle to understand how to approach problems independently. That has started affecting my confidence and motivation.
Has anyone made a similar transition from internal IT or systems administration into DevOps or platform engineering? What would be a good way to build the fundamentals, find useful projects, and use AI without becoming dependent on it?
4 Answers
Being the only platform engineer without a senior peer to learn from is probably the most difficult part of this situation. Talk with the application and service owners and ask them where their current processes or infrastructure cause friction. Treat those engineers as your customers and turn their pain points into small platform improvements. That gives you practical projects while also helping you understand the systems you support.
Your previous systems administration experience is still valuable, but platform engineering usually requires stronger depth in areas such as version control, scripting, infrastructure as code, CI/CD, containers, networking, observability, cloud permissions, and reliability. Build small labs around those fundamentals and document what you learn. Also review any changes suggested by AI carefully and test them in a safe environment. Depending on generated answers without understanding them can become risky when something fails in production.
Three months is not a long time when you have moved into a substantially different role. AI can be useful, but do not just copy its solution and move on. Ask it to explain the reasoning, what could fail, and why each component is needed. Then rebuild a small version yourself without assistance. Use it as a tutor or source of hints rather than an autopilot. Progress is not necessarily being able to do everything unaided; it is understanding more of the suggested solution each month.
Talk openly with your manager about what success looks like for the role and what areas you are expected to own. Then speak with the developers and application owners about recurring problems such as deployment friction, unclear documentation, unreliable monitoring, access controls, excessive manual work, or confusing infrastructure workflows. Pick one contained problem at a time and learn the underlying networking, operating system, cloud, security, and software delivery concepts as you solve it.
Related Questions
Can't Load PhpMyadmin On After Server Update
Redirect www to non-www in Apache Conf
How To Check If Your SSL Cert Is SHA 1
Windows TrackPad Gestures