Code Story: Insights from Startup Tech Leaders

Developer Chats - Petr Petrenko

Dec 17, 2025

Ask episode

Chapters

Transcript

Episode notes

Today, we are continuing our series, entitled Developer Chats - hearing from the large scale system builders themselves.

In this episode, we are talking with Petr Petrenko, Senior PHP Backend Engineer. Petr will take us through his developer journey, in working on large scale backends, managing the tension between stability and innovation, and designing systems to interact with culturally different economies.

Questions

You’ve worked on large-scale backends that serve millions of users. At what point do systems start to outgrow the teams that built them?
At some point, every mature backend reaches a stage where rewriting is no longer realistic. How do you recognize when a system has crossed that line, and what’s the right way to handle it?
There’s always this tension between stability and innovation. How do you decide when a system needs refactoring versus when you just need to live with the technical debt?
Let’s talk about the human side of legacy systems — what have you learned about culture, documentation, and knowledge transfer that keeps old systems alive and reliable?
You’ve also built and maintained complex payment systems for global users. What’s something most engineers underestimate about cross-border transactions?
When you’re designing systems that deal with different currencies, laws, and tax regulations, how do you balance the technical with the ethical — for example, user privacy or data sovereignty?
For engineers listening who want to build something durable — not just fast — what advice would you give about writing code that will still make sense years from now?
One of your most impressive projects is a high-performance image-matching system you built yourself, capable of scanning tens of millions of images with sub-second results. Can you walk us through the moment you realized you needed to redesign the system — and what engineering choices made that level of performance possible?
You’ve also worked on billing systems and fraud mitigation at scale. Was there ever a moment when you had to choose between a technically “clean” solution and a solution that better protected users or the business? How did you make that call?