I have been a developer for most of my conscious life, starting in times when Gigabytes were still puny little kilobytes, and all disks were very floppy indeed. Over those years, I've designed and developed a lot of web and desktop systems, ranging from inventory management and POS systems to high-performance web applications such as njuskalo.hr, and I’ve also been one of the founders and have served as CTO of one of the more prominent Croatian Web agencies in the late-2010s - called Trikoder. As a consequence, during my career, I had the opportunity to “play” every conceivable role (except that of a graphics designer - I’m terrible at that) within the software engineering process. The result of that experience was that I had ample opportunities to see and feel firsthand which approaches worked in real life and, more importantly, which ones did not, and over time figure out why that was so. Currently, I am heavily involved in the quickly emerging field of AI-assisted software engineering, and especially how the already established, robust, and transparent software engineering processes and practices in a given organization can be "supercharged" by LLMs and AI agents without introducing that dreaded "AI slop" into your mission-critical enterprise systems.