We have started a series of interviews with Ukrainian tech professionals, which put a light on the specific fields of the tech sphere in Ukraine, share experience and advice to the startups, and explore the opportunities Ukraine has to offer in the global economy.
The first person we talk to is a high-level AI specialist Borys Pratsiuk, Chief Technology Officer of Scalarr – Machine Learning-based fraud detection tool.
TechUkraine: When and how did you start your path in data science, AI, and machine learning?
Borys: My scientific career started in 2007 when I decided to become a Ph.D. student at NTUU ‘KPI” and started an investigation in the area of solid-state electronica. In 2012, I got my Ph.D. diploma and put science on hold. Then, in 2015, I went back to the research field when I became Head of R&D in Ciklum. Now, I’m a CTO at Scalarr where we use machine learning and neural networks to process client’s data and get insights out of it.
TechUkraine: You lived and worked for 5 years in South Korea. What are the most significant insights and professional experiences you gained while being there?
Borys: Actually, I only lived 9 months in South Korea, and the rest of the time I worked remotely from Ukraine; but I did learn a lot in South Korea. I had the opportunity to work in the R&D department where my main task was to be in charge of innovation in the Bluetooth and communication protocols area. I worked on several initiatives to solve real signal processing tasks on DSPs and worked on a lot of exciting and challenging projects dealing with embedded microprocessors. As a result of my first 6 months of engineering, I built a prototype for a wearable Bluetooth sensor. And in 2007, this was very innovative since the IoT hype and the Apple watch era hadn’t started just yet.
But the main learning was how organized the R&D processes were and what you could do to drive innovation.
TechUkraine: Fraud and anti-fraud systems: How does Scalarr help clients? Which markets and industries are the most active in using the technologies you provide?
Borys: Since 2016, Scalarr has fought ad fraud in some of the most challenging times as the digital ads ecosystem if flooded with fraudsters’ schemes and arbitrariness. Bad actors generated more than 40% of mimicking ads which were bought by clients. As a result, the industry’s losses reached up to $1 Billion per year. In fact, our experts predict losses up to $16b in 2020, which comes to show how fraudsters have evolved over the last 4 years, with a 16x increase in losses.
That’s why Scalarr’s mission is to build a transparent mobile ad market, free of fraud. We monitor mobile ad traffic to detect mobile ad fraud by unearthing fraudulent patterns in data with the use of powerful ML models and neural networks, thus safeguarding the marketing budgets of our clients.
The most affected spheres by fraud are:
- Finance apps
TechUkraine: Where is Ukraine on the international map of AI, data science, machine learning? What should be done to achieve optimal results and to compete with key players – the US and China?
Borys: We are just at the beginning of the global AI competition. To our advantage, we have brilliant scientists who are actively involved in different research projects with Machine Learning and Deep Learning technologies. But without solid government support in both education and business areas, we will find it very hard to become leaders in the industry.
Why should the state be the customer of AI transformation? Let’s take a look at the examples of the countries that have succeeded. For example, China has invested billions of dollars in training and universities for educating millions of neural network engineers. It also built the most powerful video surveillance network that generates enormous datasets. According to this data, research institutes and companies have built AI algorithms for medicine, agro-sector, security, and many other spheres.
Top 5 priorities for the Ukrainian government:
- Establish the new University of AI where the lectors are experienced insiders of the sphere and invite international professionals
- Complete reengineering of the educational system in all universities and provide up-to-date infrastructure and an appropriate environment to study (creating an atmosphere where the students want to study, not just waste time).
- Develop a digital infostructure with open data for universities and students.
- Grant support to students with top marks and who aim to start their own AI-based business.
- Provide security for business processes and simplify the establishment of entrepreneurs (a new group for individual entrepreneurs – free of any pressure).
TechUkraine: Why is AI the most demanded and the most-hyped technology in this pandemic and how can it solve global challenges in healthcare, education, and security? Could neural networks predict the next big crises?
Borys: So-called Artificial Intelligence (AI) is a set of mathematical formulas and algorithms that have been applied to a massive amount of training data. As a result, certain rules have been formed by which the algorithm makes a decision. Therefore, it can be used to solve many problems.
For example, you can automatically track the quality of the seedlings in harvest fields and sow the grain in the early stages, which can reduce the yield. Or monitor the health of animals and their disease rates to positively affect the end quality of food.
Also, AI can automatically detect MRI signals and look for tumors, which can directly save lives.
The main advantage of neural networks is that they can take millions of metrics and make decisions. Therefore, it is up to us whether people want to invest in such technologies and build an algorithm that will track and detect everything.
TechUkraine: Recommendation for startups: how to find out a niche on the market, develop the company, and stay strong/more reliable in the crises?
Borys: The recommendation is straightforward, starting with the problem. Communicate with different businesses on different online meet-ups, grow your networking, ask about growing pains in the industry. And then, you will find what is worth your time.
If you have an idea, start with a very fast prototype/MVP that will solve the problem. Remove all the extra stuff (design, advertising, marketing, additional features for convenience, etc.) And to survive the crisis, do cost optimization. You can even work remotely for 1 year (many famous startups have started out in a garage).
TechUkraine: What are the worldwide programs that support AI startups during COVID? How can startups get additional funding worldwide? (list of reading resources)
Borys: In all honesty, I don’t know of any program that fits the bill. More then ever, investors are now paying attention to businesses that operate without the binding of a physical location and who continue to work during the quarantine. Therefore, they continue to invest in digital solutions and platforms with pleasure. Or, they offer quarantine loans to companies they invested money in to help them survive the crisis.