The architect who said he never left engineering
For the best part of ten years I have described myself, more or less unprompted, as the architect who never really left engineering. In interviews, in team introductions, I used a throwaway line to put people at ease. I still deploy infrastructure, I still know my way around a terminal, I'm not one of those architects who draws boxes and disappears. I said it often enough that I believed it. Then…
For nearly a decade, the author consistently introduced themselves as the architect who never truly abandoned engineering, using this phrase to ease people's perceptions. In discussions and introductions, they maintained a hands-on approach in infrastructure, confidently navigating through terminals. However, a project's need for additional AWS landing zone help led to a shift in perspective.
After a few months of involvement, the author discovered a less favorable view of their distance from practical engineering. The task at hand required more technical expertise than the team possessed, which the author possessed in AWS and on paper. Despite being asked to build control tower guardrails, account vending, and plumbing for a multi-account AWS estate, the author hesitated due to a lack of confidence in their technical abilities.
Initially, they turned to AI coding assistants, expecting a quick return to their previous fluency. However, the assistants produced code that did not accurately address the given problem. The author discovered that these tools required a deeper understanding of the context, which they lacked. Using Amazon Q proved more successful when they forced the assistant to plan before coding.
This approach helped maintain discipline in coding conventions and naming patterns. The author found that their proficiency in understanding the codebase was the missing link in their engineering capabilities. While AI coding assistants improved over time, the author's own understanding of the code remained a limitation. Despite this, they recognized that infrastructure as code had evolved into real software engineering, with practices such as testing, modularity, and code review becoming more prevalent.
The author concluded that AI assistants are not always the best solution for infrastructure tasks, concluding that the real improvement came from the author's own growth and adaptation to the changing landscape of engineering.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
