About the Project
'HSE University's Age-Mates'
2022 marks the 30th anniversary of the founding of HSE University. Many of the university’s peers—those born in 1992—now work and study here. Thirty-year-old HSE graduates work in various fields, from business and fintech to IT and contemporary art. As part of the new ‘HSE University's Age-Mates’ project, some of them have shared their stories and talked about what they like about the university.
Alexey Masyutin is Russia’s first research doctorate in computer science, having earned his PhD in Computer Science from HSE University in 2018. He currently heads the HSE AI Centre and oversees the joint master’s programme between HSE University and Sber. In this interview with Age-Mates, he described when HSE University first came into his life and how it is helping him to make his dreams come true.
When did HSE University first come into your life?
In 2005. In the 8th grade, I went to the Erudite Club on Malaya Ordynka and, from the 9th grade on, I attended classes in the HSE Faculty of Pre-University Training where we focused specifically on passing the entrance exams. I really liked math, and at one point, I even wanted to attend the Faculty of Mechanics and Mathematics at Moscow State University. However, applicants had a good chance of getting in only if they could solve four test problems out of six, whereas I was certain of solving only three. HSE University is also strong in mathematics, and besides, I liked how it applied mathematics to real life. So, in the end, I decided to study at HSE University.
Why did you choose the Faculty of Economic Sciences?
HSE University was still gaining a name for itself and people kept asking, ‘Which school are you going to?’ I considered ICEF and FES to be the strongest options. The ICEF programme was expensive and did not offer financial aid. However, I wanted—and managed—to get a scholarship with FES.
Where did you want to work after graduation?
In the industry. My first job was with TechnoNICOL, a domestic manufacturer of building materials. They needed to forecast demand in order to plan distribution and production, so they hired three interns, including me, to build the models. We took sales statistics and added general data such as average market prices. Then I worked in the banking sector. I set up risk reporting at RN Bank and for the Life Financial Group I built models showing which customers we should offer settlement and cash services, acquiring or credit. I really liked it. Some time later, I moved to the validation department at Sber.
In 2017, you became head of the HSE University and Sber joint master’s programme in Financial Technologies and Data Analysis at the Faculty of Computer Science. How did that come into existence?
The so-called AI transformation has begun at Sber. This is where routine processes are automated through the analysis of previous data and forecasting. For example, dispatchers in IT support used to direct tasks to certain departments, but now a computer program does this, automatically choosing the necessary department based on the content. Now the dispatcher can get additional training to handle less routine, more substantial tasks.
Actually, all banks do this, but Sber is of such an unprecedented scale that a huge number of specialists were needed who could translate these operations into models. That’s why Herman Gref turned to Yaroslav Kuzminov with a proposal to start a master’s programme—and it has been going strong now for five years.
Why were you invited to run the master’s programme?
They needed a person who knew the needs and goals of Sber and also understood how the university works. I was just finishing graduate school at FES and writing a dissertation on data analysis in the banking sector. It went hand in hand with my job and I obtained some of the results from the Sber data. So they offered me the master’s project.
The most difficult part was getting started, figuring out who would teach which courses, working everything out with the instructors, conducting interviews and recruiting students. In 2017, we accepted 30 students; in 2019, it was up to 45, and in 2020, there were almost 60. After two years, word of mouth started working . The programme is thriving now and continues to develop thanks to the community of alumni. And last year I received a new offer—to head the AI Centre at HSE University.
How did that come into being?
HSE University won a competition in 2021 and became one of six Russian universities where such institutions were established. Again, in me they saw a person with industry experience and an understanding of how the university works who could run the programme. There are research scientists who have their own outlook and scientific interests, and there are companies that need to solve business problems; their thinking is more down to earth because they need concrete results. And I am the link between these two worlds. I started this job on April 21 but also remain responsible for overseeing the master’s programme. We are planning to have synergy, with the master’s students getting involved in the Centre’s projects, offering their solutions and, for example, defending them as their master’s theses.
How is the AI Centre organised?
This is a very large and complex project. It has various areas of focus and employs 340 people. I interact with the leaders of all of the Centre’s 25 projects. We also have a project management office.
Here, as in the industry, there is no literal concept of ‘subordinates’, of giving an order that everyone must then go and carry out. Everything is more flexible at the university. Each project manager has their own name, vision, roadmap and research. In addition to their work at the Centre, they have their own duties such as teaching or helping graduate students prepare to defend their theses. At the Centre, each is responsible for their particular part of the work and for implementing the KPIs.
Who are the AI Centre’s main partners?
Sber, Yandex and MTS. For Sber, we are improving the virtual assistants of its Salyut family and working with so-called language models that, for example, tell the machine what to answer and in which situation. We are also solving several problems for the SberDevices division to improve their Jazz product. This is similar to Zoom but has additional cool features like automatic call transcription.
For Yandex, we make programs that allow us to predict the trajectory of objects. This is for driverless vehicles. When people drive cars, they understand when the other drivers will change lanes and, in advance, act reflexively based on how they anticipate the actions of other people on the road. We are improving drone models to make driving safer.
And for MTS, we make models that improve sound and image quality when transmitting over an overloaded channel. In addition, we are constantly looking for new partners in industry. I meet with companies to discuss how they can apply what we’re developing.
Here are two interesting examples. The first is software that allows drug developers to make a short list of molecular shapes that must come into contact with the antigen. This software makes it possible to reduce the number of physical experiments needed to test a drug compound. In this way, you can reduce the cost and time of finding a formula that actually blocks an antigen. This bio-informatics development from the FCS is of interest to BIOCAD. For another company, Genotek, we will be developing AI models to evaluate polygenic risk scores for cardiovascular disease. In addition, the International Bioinformatics Laboratory of HSE University plans to develop AI systems in cooperation with Genotek to help people search for distant relatives use genotyping results.
In a second example, my colleagues at FES are working on a way to add analytics based on open sources to classical financial indicators to calculate a sort of index of the mood on financial markets. This would enable the Central Bank and the stock exchange to offer more effective regulatory measures during periods of high market volatility, at times when they must switch to an alternative trading method.
Has HSE always supported your ideas?
I suppose so, yes. For example, when we started the master’s programme, we held hackathons along with other companies and HSE University allocated a prize fund. Our co-workers did a lot of legal work for us. For the summer school, we were given both a platform and employees who organised and moderated everything. A second summer school was held in late May of this year. I am very grateful to the university for all of this.
What is your dream?
I wish there were no gap between the academic world and industry so that they could enrich each other. Many companies are doing this: Yandex, Sber and Tinkoff. I want to contribute also, and the AI Centre’s solutions, that find application in industry, help make my dream a reality.
What would you wish for HSE University on its 30th anniversary?
There are many teams working at the university that have a vision of what they can do for the country, for companies and for science. It is my wish that such teams would always receive the greatest possible support—both financial and administrative—so as to overcome every difficulty and realise their dreams. It’s like in the film that my HSE co-worker, Alexey Naumov, recommended to me—Taming of the Fire—that depicts the life of Sergey Korolev. In today’s terminology, Korolev was also constantly receiving state investments and grants and applying them. Some things worked out and some didn’t, but he was always striving and full of passion. He managed to find the necessary support and eventually achieved monumental results.