What is the Balance — AI Assisted Coding in Network Automation
A fellow grad student, Ayush Mishra, and I started the early stages of our careers learning programming before AI-assisted coding agents became mainstream. We were recently discussing how, in the past, a major part of career growth came from deepening our software knowledge and sharpening our programming skills. While these are still absolutely important, the latter is clearly evolving.
Instead of spending hours figuring out how to correctly implement a data structure, sifting through Stack Overflow, or shadowboxing demons in the code, that time is now often spent reviewing what AI has written and verifying its correctness.
It’s difficult to describe, but we both can’t seem to shake a subtle feeling of guilt. The old process—defining a problem, designing an algorithm, implementing the code, and finally seeing it work—has shifted toward defining a problem, “vibe coding” the algorithms, code, and tests, and then seeing it work. At times, it feels like I’m cheating myself out of the hard-earned knowledge and experience that used to come from the programming portion of the work.
Am I just stuck in the AI matrix? Engineering tools and methods have always evolved, and I’m happily participating in that evolution. Still, it feels like systems design knowledge is rapidly becoming a major competitive edge—perhaps even more so than before.
We wanted to ask our peers and industry veterans: how should we position ourselves alongside AI to stay healthy and competitive in network engineering roles? What is the industry trending toward? Should we be fully embracing AI-assisted tools and shifting more focus toward systems design?