Last time, we introduced the engineers behind the UI. Today, we’re moving deeper into the platform – to the backend, the part that keeps the heart of flespi beating...
Lead Software Engineer
I’ve been working at Gurtam since 2008. Back then, the company had no more than ten people, and with that small team we built Wialon – which is still the flagship product today.
A few years later, in 2015, a new direction was launched, and several experienced engineers were assigned to it. That’s how flespi came to life – as a telematics backend service for different platforms, including competitors of Wialon.
From a technical perspective, the flespi backend can be split into two main parts: the database with analytics and the data processing service for telematics devices. I’m involved in both, but most of my work is focused on the services that handle internal communication between these components, as well as external access for clients.
As for technologies, the core of Wialon was written in C++, while flespi is built in C – so that’s the language I feel closest to right now.
What I really enjoy about my work is that we’re still building a large and complex product with a relatively small team.
At the same time, everyone genuinely cares. We’re not afraid to argue, suggest ideas, and have real discussions about important topics – while still respecting each other’s opinions.
That said, there are parts of the job that are more stressful than others. For example, rolling out complex features is always a bit nerve-wracking – there’s always a chance something might unintentionally affect the system and impact our clients’ projects.
When it comes to AI tools, I honestly can’t imagine my work without them anymore. I prefer using them in a semi-controlled way – guiding them in the direction I need. Of course, everything still has to be checked. There’s no full trust yet, and there won’t be until everything is properly covered by tests – both input and output, just like when working with another engineer.
At the same time, once you start using AI, there’s really no going back. Even short outages – when services like OpenAI or Anthropic go down – you immediately feel how much you rely on them. And this dependency will only grow. So the real question is not whether to use AI, but how to adapt to it.
In my view, concerns about AI replacing learning mostly apply to beginners. When you’ve been working in this field for years, a large part of your job becomes routine – and AI helps eliminate that. You already understand how things should work, you can clearly describe it, and AI takes care of the implementation. It removes the most exhausting part of the job, which makes you more productive and helps avoid burnout.
Outside of work, my main focus is my family and computer games. I’ve been gaming since the early 90s, and I still enjoy it just as much. Any genre really – shooters, strategy, you name it. If it’s a PC game from the last 30 years, chances are I can hold a conversation about it. That’s my way to unwind.
Speaking of stress, one of the most intense moments is releasing updates. You push a new version, and there’s always that tension – what if something goes wrong? But then some time passes – half a day, maybe a day – and you see that everything is stable. That feeling of relief when you realize nothing broke, and everything is running smoothly, is incredibly satisfying. In our field, that’s what matters most – for clients, the best outcome is when nothing breaks.
And if I could say something to flespi users, it would be simple: stay with us. We’re evolving – and we’re moving even faster now than we did in the early days.
Lead Software Engineer
I joined the flespi team right from the start in 2015, when it was still more of a development lab. Before that, I worked on Wialon – first in the Hardware Support team starting in 2011, and later in Automation and WDC (Wialon Data Center) teams.
At flespi, I work across several areas – mostly backend, with a bit of system administration as well. One of my main focuses is the Telematics Gateway, especially its parsing part – the PVM compiler and runtime. I’m also involved in the media data processing service.
On top of that, I deal with system-level components like integrating the Lua scripting engine into our services. Sometimes I handle infrastructure and DevOps-related things – how services are deployed, internal tools, that kind of stuff. Part of system administration is also on me – for example, making sure PostgreSQL clusters are properly set up and running.
In general, I try to apply my experience wherever it’s needed. And yes, I occasionally try to push our “dinosaur” part of the team a bit closer to modern development practices. :)
Another part of my role is incident response – basically 24/7. Luckily, the system has been quite stable in recent years, so interruptions outside working hours are rare. Most of the time, it’s something minor – like a database disk failure – the kind of thing users usually don't even notice. So overall, I guess I’m a bit of a jack-of-all-trades when it comes to backend and system stuff.
When it comes to programming languages, I’m comfortable with Lua, C, Python, JavaScript, TypeScript, and Bash. In general, I see programming languages as tools for expressing ideas – and different ideas fit different tools.
A programming language is just a way to express your thoughts, and different thoughts require different tools.
As long as it’s a well-established language, I’m fine with it. I’ve even built a few of my own along the way – like PVM. What I enjoy most, though, is the variety. There are tasks from completely different domains, and a lot of freedom in how you approach them. The team is strong – there’s always something to learn from any discussion we have.
There’s also space for growth – and even a bit of perfectionism. Sometimes you solve something quickly, and that feels great. Sometimes you spend a long time, but the result is solid – something you can be proud of. And sometimes you really have to struggle – but then comes that moment when everything clicks.
That “Eureka” moment, when you finally come up with a clean solution – that’s probably the best part.
There’s always something to improve. In a world that’s naturally chaotic, parts of our system actually look very solid – but there’s always something to refine. On the flip side, with such a wide range of responsibilities, planning is tricky. Every day, something comes up that needs immediate attention, so priorities shift and you constantly have to switch context. That’s probably the most frustrating part. :)
It’s also rare to get a long, uninterrupted stretch of time to really dive into something. Working directly with customers isn’t really my thing either – although it’s still nice to know that something I built helped someone. And of course, there’s always some routine and a bit of bureaucracy, even if not too much. The downside of having a lot of different tasks is that many of them are urgent. Work-life balance, to be honest, is probably a bit shifted toward work.
Who can help? As for AI, it feels like a new stage for all of humankind. Things won’t be the same as before. I was quite skeptical at first – mostly because of inconsistent results. And honestly, “AI” still sounds more like marketing than reality.
Over time, I started treating it as just another tool – a very powerful one, if used right and in the right context. But you still need to verify everything. Always.
It’s not about whether to use AI – it’s about how to use it properly.
It definitely changes the way I work. And overall, it makes things more interesting. Right now, we’re all just learning how to use it and where it fits best.
Outside of work, I enjoy spending time with my family, traveling, and reading science fiction books. I’m passionate about learning new tools and frameworks, and exploring emerging technologies. As for stressful situations, there’s no special ritual. Unexpected things are just part of the job and life itself. In the office, there’s always something else going on, so switching between tasks kind of resets you automatically.
If I could say something to flespi users – I genuinely enjoy working on this product and seeing how differently people use it. So go ahead – explore it, experiment with it. It makes things more interesting – for you and for us.
Well, there are only 2 guys left, and yes, the last post in the series will introduce Aliaksei Shchurko, Chief Gurtam Architect and CEO, together with Jan Bartnitsky, a Software Engineer responsible for business development processes at flespi. Stay tuned!