I have about four years of IT experience, with most of it focused on DevOps and some vulnerability remediation. I started on a traditional technology team, then moved into a DevOps role where I worked on CI/CD pipelines, infrastructure as code, and related automation. I was able to learn quickly and deliver projects, even though the rest of my team consisted mostly of architects and technical leads with decades of experience.
After a project ended and some senior colleagues moved on, I became the CI/CD architect for a new initiative. I designed and implemented the delivery pipelines and infrastructure, but I do not know much about the product's business logic, testing, or application internals. I am often included in architecture meetings, but I struggle to follow parts of the discussion.
I also worry that my pipeline and infrastructure code is not well structured and may be missing basic best practices. Being surrounded by much more experienced people has made me feel intimidated and as though I do not know enough. Recently, I have started relying heavily on AI to write code, and I am concerned that I am no longer developing my own problem-solving skills.
Is this level of uncertainty normal at this stage? How can I use AI responsibly, strengthen my fundamentals, and build the skills and confidence needed to keep up with the current DevOps market?
2 Answers
Broaden your understanding gradually instead of trying to become an application architect overnight. Learn the product's basic business purpose, deployment flow, dependencies, observability, security controls, and testing strategy. Ask teammates to explain the architecture in practical terms and keep your own notes.
For your DevOps work, review fundamentals such as version control, CI/CD design, infrastructure-as-code structure, secrets management, testing, rollback strategy, monitoring, and cloud security. Build a small portfolio project or lab where you can practice these concepts end to end. Consistent hands-on repetition will improve both your skills and your confidence.
Feeling out of place around people with 20-plus years of experience is completely normal. The important thing is to focus on understanding the reasoning behind the work rather than trying to memorize every tool or produce code without help. The "why" matters more than the exact syntax.
AI can be useful, but treat its output as a draft. Ask it to explain each part, point you to official documentation, identify security and reliability risks, and suggest tests. Then review and modify the result yourself. Anything delivered under your name should be something you can explain and maintain.
A small home lab can help a lot. Build a simple application and create the pipeline and infrastructure yourself, then break and improve it. That gives you space to learn without the pressure of a production deadline.