Inside Exadel’s Data Department: How We Work, Solve, and Learn

A Glimpse of Working in Exadel’s Data Department

You open your inbox. There’s a photo of a crumpled invoice, a half-filled spreadsheet, a few PDFs, a PowerPoint screenshot, something that might’ve once been a scanned napkin…
The point is: there are a lot of them. With a capital L. All of it is technically “data.” All of it is now your problem. You need to make it searchable, structured, and reliable; whatever you build, it has to work across formats, versions, and thousands of incoming files.

“Sounds like a job for a human, not software. Right?” – says Dzianis Reutski, Head of Exadel’s Data Processing & Analysis Department, kicking off an insightful conversation about data, people, and systems.

“Well, it’s the kind of problem we’ve had to solve, and GenAI turned out to be the right tool for it.”

Turns out that the project wasn’t an exception but a glimpse into how our Data team operates. So we sat down with Dzianis Reutski to get behind the curtain of the stories we’ve heard, and what we found confirmed what we suspected: This is not your average data department. Let us show you why.

Systems & People

Before leading Exadel’s Data Processing & Analysis Department, Dzianis was an engineer—and then an architect. That career path, he says, taught him two things:

“I love systems. And I care deeply about people. It’s the intersection of building structured, working systems and understanding the people behind them.
For me, it’s all about learning what drives someone, where they shine, where they struggle—and building an environment where their best qualities can actually grow. That’s the part I enjoy most.”

That philosophy isn’t just Dzianis’s personal leadership style and it’s visible in how the department functions day to day. When asked to describe his team, he doesn’t give a mission statement. He gives three lines:

Then he invited us into the world behind those three sentences—where every decision is driven by evidence, logic, and questions that demand better answers.

“We bring structure to chaos”

That’s how Dzianis sums up the work of Data Department—a space where data engineers and data scientists collaborate tightly, rather than function as separate silos.

“In some sense, modern Data Scientists aren’t that different from ancient oracles. Back then, it was bird entrails — now it’s dataframes. We throw messy inputs into statistical models, look for structure in the noise, and try to figure out what’s really going on. Not necessarily to forecast the future, but often just to understand the present, which, frankly, is difficult enough.”

Understanding the present—especially in fragmented, undocumented systems—takes more than talent. Without trust, it falls apart. This team has earned that trust. Trusted to advise, not just deliver. “Our clients typically give us a fair degree of technical trust, which allows room for unconventional decisions. Sometimes the ask is as vague as: “Can you make this work? Like, press-a-button kind of work? In those cases, experimentation becomes a requirement.”

That culture of experimentation lives at the heart of the department and across Exadel’s Practices (R&D).

“Data is still a young domain. There aren’t that many paved roads. But there’s plenty of science, precision, and trial-and-error included.”

Inside Exadel’s internal Practices— especially AI and Platform Modernization — engineers test new tools and ideas without constraints: no ticket backlog, no deadline pressure. Just like-minded people, ideas worth chasing, and the space to chase them.

“Ideas are tested and filtered. The goal is clear: to capture our collective engineering experience and amplify it with modern tooling and methods.”

The biggest catalyst for that experimentation? AI.

“AI is no longer a side experiment.” As Dzianis puts it, at Exadel, AI is now steadily embedded into engineering workflows. Incremental gains at every step of the software development lifecycle are adding up to something more powerful. “Some of the most promising directions so far include assisting non-technical roles like BAs and POs, reducing QA routine overhead, and exploring code generation for tests and navigation support across complex codebases.” But none of this freedom is taken for granted. “There’s flexibility but it’s earned, applied with context, and grounded in engineering judgment.”

Once upon a time… A vendor disappeared

“There used to be a saying in our engineering circles: We show up when your vendor gives up.”

One of our clients, a global retailer with thousands of branches, needed a financial reporting system. Taxes, returns, subscriptions, point-of-sale operations — a lot of it. They hired a vendor, paid the budget, and got something in return. Technically, it existed. Practically, it didn’t work. Then the vendor disappeared. Along with the documentation. What was left: a half-integrated, mostly broken system, a looming regulatory deadline, and no way to restart from scratch. Our team was asked to step in, figure it out, and make it work fast. We reverse-engineered what we could, patched what couldn’t be rebuilt, and rebuilt what couldn’t be patched.

Under time pressure, with incomplete information, we delivered a working system that’s now in production, running daily across thousands of data points and jurisdictions.

It’s a working tool — not flashy, but effective. It enables branches to operate without disruption, generates reliable financial reports, and doesn’t block business operations.

Legacy Migration: The Nightmare That Won’t Die

It turns out not every challenge in the department is about building something new. Sometimes, it’s about understanding something old and haunted.

“What’s an engineer’s worst nightmare? Depends on who you ask. But for me, it’s migrating a legacy system to modern tech. Sure, moving from Java 8 to 17 is painful. Going from ‘old Angular’ to ‘new Angular’? Not fun either. But try migrating something written in Visual Basic — and no, I don’t mean VB.NET. I mean actual VB.

Or worse, some mainframe-only language that hasn’t seen a stable release since the 1980s. The last person who understood it retired five years ago. The system? It still runs. It’s probably huge. And it’s probably critical. Exadel has dealt with these things. And trust me — it hurts. We’re talking months or years of reverse engineering. You need to find a software necromancer fluent in forgotten syntax just to read the code.”

So what do we do? We invent a whole new tool that speaks the language and makes the legacy nightmare go quiet. Welcome to the world, Exadel Lumea.

Learning: Not Accidental. Expected.

Thinking about joining? Dzianis shares an unofficial cheat code. The moment that makes us go “yes, you belong here” is when we spot a genuinely curious mind. The kind that fearlessly keeps asking, exploring, and earning. Because, when it comes to learning, Dzianis is clear: Exadel does it right.

“Exadel has a separate department for Learning & Development, and this says a lot. Honestly, it’s one of the most complete and scalable learning ecosystems I’ve seen in an IT company.
Thanks to our amazing team, learning here is structured and well-run — not just a random list of training links or a Slack thread with ‘cool stuff.’”

The “Learn and Mentor” principle is everywhere—supported by role-based courses, bootcamps, workshops, knowledge-sharing communities, sandbox sessions, and on-demand training. The resources are there. Self-learning doesn’t take a backseat here. Quite the opposite: it’s expected, encouraged, and backed by structure.

“When you deliberately hire people for their intelligence, curiosity, and technical depth — and then give them space to grow — you shouldn’t be surprised when they actually want to keep learning. The key is to create the right environment and not get in their way.”

Inside the Data Department, knowledge sharing takes multiple forms. There are three active, and very much alive, Engineering Communities for AI, Data Science, and Data Engineering.
These aren’t ghost groups: people show up, swap tools, share lessons from real projects, and push each other — and the department — forward.

More Unofficial Cheat Codes for Passing Exadel’s Data Interview

There are even more unofficial cheat codes, and Dzianis was generous enough to share them during our talk. Turns out, there’s no such thing as a great data oracle without these three:

This combination — curiosity, technical depth, precision, resilience, and a sense of humor — is what makes Exadel’s engineering culture world-class.