One of the most famous Czech AI companies, FF Group, was founded in 2008 as a Kyiv-based business. Initially, it was an IT integrator, but then the team began to focus on development and implementation: it carried out projects for Ukrhydroenergo and shopping centers, developed software and video surveillance systems, and attempted to enter other countries with its software.
"But we were atypical developers, and we remain so today. The company's founders come from scientific and science-related backgrounds. I myself worked in a scientific environment," said Alexander Osipov, founder and CEO of FF Group, in an interview with 
.
"Formally, we work in the IT market, but the core of the company is physics and mathematics. That is, we work one step ahead of the algorithms. We have learned to build bridges between science and implementation. We worked, gained knowledge, people, and talent at the Institute of Cybernetics, the Faculty of Mechanics and Mathematics, and the Faculty of Radiophysics at Kyiv University. We learned how to turn this into working code," he explained. This continued until 2016-2017, when the story of moving the business to Prague began.
How did you end up in the Czech Republic?
– To answer briefly, it was by accident. But in more detail, I love downhill skiing and often traveled to Slovakia. I fell in love with the beauty of this country and began to look at it from all angles, including from a business perspective. Then we met a professional in the field of video camera production who worked for one of the leading manufacturing companies in Prague. And that's when everything came together — the friendliness, the safety of the country, the good conditions for scaling the business, and working with camera manufacturing professionals. The latter was especially important, as FF Group was entering a new era — Edge recognition.
Alexander Osipov
I thought: we can keep our R&D core (research — ) and distribution network in Ukraine, and scale up in Europe and further around the world by opening an office in Prague. The time had simply come when we needed to monetize our ideas and bring the product to market: actively test it, look for partners, and sell through their channels. Unfortunately, being only in Ukraine, this is impossible — such is the reality of the market. Just as it is impossible to fully conduct business in the US from Europe.
With the move to the Czech Republic, we accelerated the restructuring of our business towards sales. This was facilitated by clear, transparent financial relations, tax control and administration, as well as a generally comfortable environment for doing business. And somehow, without even noticing it, we got used to this rhythm and began to actively cooperate with European companies.
In which countries can you engage in AI
The words "AI company" and "Czech Republic" in the same sentence seem rather unusual. Can companies like yours operate from anywhere geographically?
– AI can be done in different markets, although, of course, there are some peculiarities. AI can also be very different. From voice recognition to the popular GPT, but at its core is good old mathematics, which does not change as quickly as fashion trends and applications. If you really want to do AI, you need a school. You need people with mathematics. Roughly speaking, those who can multiply two matrices in their head. Otherwise, you will simply be a user of other people's developments.
To strengthen R&D and the production core, we are bringing people with a strong background in the exact sciences and programming into the team. These are mainly talented specialists from Ukraine.
Your company's website features the slogan "Engineering on Edge." What does that mean?
– License plate recognition was already a business 10 years before we started doing it. But video signal processing was done on server platforms.
We were among the first in the world to transfer heavy, complex video analysis algorithms to ordinary commercial (non-specialized) cameras — that is, to the front end, or edge. In this way, we brought recognition to the mass market, where such cameras became available even for residential complexes, and the police and retail sector were able to scale their projects.
What is the problem? The cameras are very compact, but we can embed complex algorithms in them. This requires very deep mathematics. How do you process millions of raw data where you don't have the resources to process it? That's what "on Edge" is all about. Roughly speaking, we rewrite the addition and multiplication operations to make it work fast.
The development of the cameras and chips themselves, which are becoming more and more powerful, also plays an important role here.
We weren't the first to do this, but we were one of the first to learn how to sell it on the mass market. Nowadays, a camera that doesn't recognize license plates is a rarity.
The word "engineering" appeared after we thought about what our marketing strength was.
We are probably not scientists, because for scientists we are very engineering-oriented, and for engineers we are very scientific. We are like guinea pigs — neither pigs nor guinea pigs. So we decided that we are probably engineering.
The inventor lives in the future
There is another marketing term – "engineering on the edge." Elon Musk is currently implementing it vividly in his Starship project, where a dozen consecutive test launches ended in explosions. There are still a year and a half to two years left before the lunar mission for which the rocket is intended, but no one has seen the engineering design of the spacecraft yet. Questions about it are answered with marketing renders. In your opinion, how does "engineering on the edge" differ from technical adventurism?
– Yes, that's a good question. I will answer as an engineer and as a techie. At the same time, I understand that there is now a political context surrounding Elon Musk; he is now not only a techie, but also a political and media phenomenon.
I consider the Tesla car a breakthrough because it changed the industry. Musk persistently promoted it, despite failures and attacks from critics.
An inventor lives in the future — he is only a guest in the present. Even the popular slogan "Fake it until you make it" reflects this idea.
I also consider myself one of those people. From the outside, it seems that we want to go to the future. No, we suffer in the present because we are already in the future.
It seems to me that Elon Musk has been on Mars for a long time, in the best sense of the word. When he draws a ship design, he really sees it. An inventor thinks not in formulas, but in pictures. He imagines that it already exists.
It starts the same way for both those who deceive and those who do — with an image. But for the latter, it's not just an image, a whole world appears in their head, and they honestly begin to live there. Not imitate, but live.
Sometimes, if you are lucky, or God kisses you on the forehead, the imaginary world becomes reality. If not, then you remain a dreamer and a fantasist. And people say that you lied, cheated, wanted to make money.
Now my friends and I are planning to create another innovation in the field of AI. I can't promise that everything will work out for us. But I can guarantee that within the framework of our agreements, I will do everything possible to make this new idea make our lives more interesting. That's how it is.
Your website states: "To stress test the DataFactory platform without code, our team created a ‘poker analytics’ product in just 12 hours." Please explain how card games are relevant here.
– Since the company is ours, we are only accountable to ourselves and our team. Sometimes interesting ideas come up that we just want to implement, show to the market, and then develop further.
Long before the big war, we had the idea of teaching the system to recognize playing cards so that casino owners could understand which players were cheating.
Of course, I'm not a fan of casinos, but BlackJack, for example, is a very cool mathematical problem. How do you determine that a player is counting cards? There are algorithms for this. They are no less complex than calculating statistics, correlations, and so on.
And we took on the task. A group of employees spent a year developing BlackJack recognition technology. We didn't sell anything, but we ended up with very powerful software. We came up with several technologies that we had never thought of before. We didn't have access to casinos to obtain data sets. So we started making synthetic sets. With only 500 photos of cards at our disposal, we were able to simulate virtually any spatial arrangement of cards on the table. How they can lie, how they overlap each other. We generated tens of millions of different combinations and trained a neural network to recognize what was happening on the table.
There were several other technologies, such as how to recognize very small objects, inscriptions on cards, with very weak cameras — this is called sub-noise recognition.
We built a set of solutions for these tasks as part of our regular business. During the COVID pandemic, when it was impossible to travel anywhere, thanks to Blackjack, we refined part of a synthetic approach based on mathematics that describes all the world's license plates.
Our team complements each other. For example, we have colleagues from the Czech Republic who are very strong in linguistics and are professional linguists, and we have colleagues from Ukraine who can surprise you with their work in mathematics. And it turns out that there is not such a big difference between linguistics and mathematics. In working on license plate recognition, we found great synergy between these groups of employees. Because license plates are also a language, albeit a very strange one. And we described it using mathematical means. And at the heart of this is the humorous BlackJack project.
A strong, tough, and frank conversation with yourself
You recently took part in a reflective game organized by Media-Kluch. Why did you need to do that?
– Thank you for not asking me what kind of game it is. I am a big fan of this technology. I started playing in 2010... I have tried to explain it many times, but one explanation is worse than the other. It is probably impossible to explain in words, but I will try again.
Technically speaking, it is analogous to quantum mechanics. After learning about it, I realized in my third year that the world is not at all what we think it is. There are no solid balls of electrons and protons. In fact, everything is different. Strictly speaking, the world does not exist.
Technologies called methodology had the same huge impact on me. Reflective games are among them.
Such games are a very strong, harsh, and frank conversation with oneself. It is a place where it is impossible to feel sorry for oneself.
In the last game, I was solving a very difficult business-social problem that I had been unable to cope with for a long time. And the game started the process.
This is what this community calls an analog model. During the game, you can build not a simulation or an imitation, but a model. Live this piece of life. And if everything was very painful and very honest, then this thing is then repeated in the real world. More precisely, it is not repeated, but continued.
First, I myself came back a another person. When I go to a game, my wife says, "Zmiy went to change his skin". That's a very apt description. There comes a time when you have to change your skin.
Second, I brought my team there. Our business is growing rapidly, and this puts a lot of pressure on the team. Many people want serious changes, but to achieve this, you need to find your inner resources, understand where to get them, and be able to share them with others. You can't demand change if you're not ready to change yourself — and that, as we know, is easier said than done.
And the game shows this very well: you either talk because you have to, or you really want to, and it comes from within you. And once you meet your true self, that's it, there's no escaping that person. Because it's you. And together, you can do a lot.
But over time, you start to drift apart from this person because of life, because of everyday routine, because of lack of money. I think I go to games and pay money just to finally be myself for 3-4 days. The enthusiasm I get usually lasts me a year, two, three.
Judging by what you say, these games are a dangerous thing.
– They can be dangerous and are not suitable for everyone. This is necessary for those who have set themselves a challenge and want to understand themselves. Who feels that something is bothering them and wants to understand how to work with it. The game is one of the methods of working with oneself. It is a powerful tool, but it must be used with care and trusted only to professionals, since psychological safety measures are required. Even with yourself, you can get into such depths... And it is important to think about whether you should go there at all?
Read also
