Before I moved into cloud engineering, I spent years in finance, audit, and procurement. It was a world governed by spreadsheets, risk mitigation, and strict compliance.
When I decided to transition into technology, I assumed my biggest hurdle would be learning the syntax. I thought I just needed to learn Python, Bash, and Linux networking, and the rest would follow.
I was wrong. The hardest part wasn’t learning the tools; it was unlearning the culture of expertise.
The Cost of Being Right
In audit, being wrong is a catastrophe. You are expected to have the answer, and that answer must be defensible against intense scrutiny. You learn to speak only when you are absolutely certain.
In engineering, especially in distributed systems, absolute certainty is a myth. The systems are too complex, the state is too dynamic, and the blast radius of a change is rarely 100% predictable.
When I first started building infrastructure, I found myself paralyzed. I was trying to audit my Terraform code before I even applied it, trying to foresee every possible failure state in a system I barely understood.
Embracing Iteration
It took months to realize that in engineering, iteration is safer than prediction.
You don’t build a resilient system by predicting every failure. You build it by assuming it will fail, and designing it to recover quickly. You test in smaller environments, you implement progressive delivery, and you rely on observability to tell you what’s actually happening, rather than relying on your own static model of the world.
Expertise in finance is about minimizing variance. Expertise in engineering is about managing it. Learning to let go of the need to be perfectly right before taking the first step was the real transition.