I’m Ihor Cherednichenko — an engineering leader and hands-on architect with 15+ years in software development, more than a decade of it spent leading multidisciplinary teams that build and run large, multi-tenant cloud platforms. These days a growing part of my work is figuring out where AI genuinely fits into the software development lifecycle — not as a demo, but as something a real team can rely on under production pressure.

This blog exists mostly out of curiosity. Agentic development is changing the methodology of building software faster than the industry can document it, and much of what gets written lands as either enthusiasm or fear. Having used these techniques daily in a large enterprise environment, I wanted a place to share what I can actually observe and measure: which practices hold up, which quietly fail, and why.

Most articles here grow out of Quorum — a research vehicle for multi-agent, semi-autonomous software development I built to study these questions in isolation, with an instrumented history and repeatable experiments instead of anecdotes.