Dmitry Sinev

I'm a software engineer and a team lead, and outside of work I build things, read a lot, and play games. What connects most of it is simple: I like understanding how things work and why they hold together — a large production system, a team, a small side project, or a good science fiction novel.

I see myself as a software engineer, broadly — not a specialist tied to one language or one kind of problem. The tools change with the job; what carries over is knowing what to build, what to check carefully, and what to leave alone. These days I work mostly on large, high-traffic web systems — online ad auctions and code that has to load quickly on millions of devices, where small delays cost real money and you can actually measure whether you got it right. Those are my favorite problems to work on: hard, unfamiliar, and possible to keep score on.

For the last few years I've also led a team, which I put together from scratch. The part I'm most proud of isn't any single hire — some of the engineers grew into strong professionals here, others were already strong when they arrived. It's that the team runs as a system that doesn't depend on me. It ships without waiting for my approval, makes good decisions when I'm not in the room, and doesn't need micromanagement to stay on track. I still review code, help with the harder trade-offs, and try to make sure what people learn stays with the team. To me, a team you have to constantly control is a team that wasn't set up well.

On the technical side, I care less about writing code quickly and more about what happens afterward. Generating code is cheap now, especially with AI in the loop. The hard and valuable parts are understanding the system, choosing trade-offs on purpose, making sure things actually work, and keeping them easy to change later. I use AI tools every day, but I check what they produce and take responsibility for whatever ships. I write more about this on the blog.

Outside work

Most of what I enjoy comes back to the same thing: systems with simple rules that grow into something complex and interesting.

I read a lot of hard science fiction — the kind that takes its science seriously. Peter Watts made me question whether consciousness is even necessary for intelligence; Heinlein left me with a stubborn optimism — the belief that one of the doors always opens into summer, if you keep trying them. The same curiosity pulls me toward evolutionary biology and toward how people actually think and make decisions.

I'm also fascinated by what AI is doing to software right now. A lot of my spare attention goes into building setups that let AI coding tools work well — giving them the right context and the right limits — and finding the point where they stop being reliable and a person needs to step back in.

I play games on purpose, not just to pass time. I like the ones with a strong point of view — roguelikes, extraction shooters, open sandboxes where the story comes out of the systems instead of a script — and I tend to avoid live-service games built mainly to keep you logging in. I also love board games, for the simple pleasure of a good ruleset and a few friends turning competitive for an evening.

My favorite project is also the least serious one. Swinobot is a Telegram bot I made for my friends and my wife: a chatbot with its own personality and memory that lives in our group chat. It remembers things you said weeks ago, joins the conversation on its own, draws pictures, and at night it "dreams" — a background job that tends and reorganizes its own memory so it can recall the right things later. What began as a way to make people laugh slowly turned into a real engineering project — a two-step pipeline that separates reasoning from voice, long-term memory, an admin panel, and a full Kubernetes setup, all far more than a chatbot strictly needs. I've enjoyed every part of it.

A few other projects run quietly in the background: this website, the setup behind it, and an over-built notes system. Away from screens, I take photographs — mostly an excuse to slow down and really look at things — lift weights, and love getting out on a bike or in the car for the ride itself.

Most of all, though, I walk. For hours, nearly every day, usually with a podcast or an audiobook playing — it's easily my favorite part of the day, and it's where a lot of my thinking happens.

Say hello

If you have a hard problem and want someone who cares about how it holds up over time, not just whether the demo works — or if you just want to talk about a science fiction book — you can reach me at me@ast.rocks.