About

What this blog is now

I’m Rizart Dokollari, a software engineer. This blog went quiet in 2021 and is back because the work changed: nearly all the code I ship is now generated by coding agents — WakaTime measured 98% over a recent seven-day period.

The change isn’t mainly speed. The bottleneck moved from writing code to everything around it, and the fast, pattern-based coding became slow, constant thinking. These posts are the lessons from working that way: teaching agents our conventions once instead of re-explaining them, building a workflow for each new bottleneck, and deciding where a human still checks in.

The 2017–2021 archive from the hand-coding era, on serverless AWS, CI/CD, and testing, stays on the site.

Who it’s for

Experienced engineers working with coding agents, or curious whether to start. Also my future self; many of these posts are notes I’ll need again.

What to expect

To the point, but not short; the flagship post runs about a thousand words. Every post opens with a TL;DR so you can take the gist and leave; the rest is the full reflection, as long as the lesson deserves, nothing padded. The substance is what working this way has actually taught me: what worked, what didn’t, and what I’d do differently.

Trade notes

I’m not writing this to broadcast. If you’re working with agents too, or further along, I want to compare: what you’ve built, where you keep humans in the loop, what you’d push back on. The comments under every post are open — or find me on GitHub or LinkedIn.