Dmitry Storcheus graduated from ICEF’s BSc programme in 2011 and then entered the MSc programme in Mathematics at The Courant Institute of Mathematical Sciences in New York (NYU). In 2017, after working for Google for several years, he continued his studies at NYU on a PhD programme created jointly with the corporation. Currently he is a Software Engineer in the Google Research department in NY and conducts research with one of the institute’s most outstanding researchers, Professor Mehryar Mohri, who is also his colleague in the Google Research laboratory. Dmitry specializes in Machine Learning (ML), in particular in AutoML – the most cutting-edge trend which allows the creation of self-studying neural networks that do not demand the participation of a programmer.
Mathematics has become your profession, although you started as an economist. Was it easy to change to technologies with your background?
It is worth noting that the mathematical training is really very good at ICEF with an emphasis on econometrics, statistics and calculus. ICEF teachers make a great contribution to the quality of studying – students really like their subjects. I studied calculus in Jeffrey Lockshin’s class; it was great and really helped pass new bars.
Truth be told, mathematics at ICEF is much stronger than on any BSc programme in Economics in the USA; Russian students are head and shoulders above western students
It wasn’t easy for me to choose the field; I have shifted focuses several times. I really liked quantitative finance, stochastic analysis, and in general I wanted to do research. I thought about following the academic track and understood that if I wanted to develop in the technical field I need to receive a degree in Mathematics. Moreover, one needs a mathematical degree to work with these subjects at the university. I started looking for a programme and university to purse a degree in Mathematics.
How did applied experience influence your choice of the engineering sphere?
I had an internship at Credit Suisse bank on an equity analyst position. I built oil companies’ models and did standard economic analyses. One needs a relevant degree to get into technologies. For example, Yandex would always prefer a programmer rather than a financier for technical tasks. During my work I grew to believe that Finance was not for me.
What about Financial technology?
To my mind, a company would still prefer a candidate with a tech degree. In management and business economics skills prevail, but the best way to enter Financial technology now is to get a tech degree after ICEF. In this case you will be a great and needed specialist. An additional education in this sphere won’t hurt, especially with the strong mathematical training that ICEF gives.
How did you evaluate this strong training when you were applying at the western university and, what is more, to the mathematical programme?
I was in many respects stronger in mathematics than my western peers, that’s a fact. But even more important was the cultural training for the global context. The ICEF educational environment itself – the way you sit exams, the articles you read, the professors you communicate with, studying in English. After ICEF it is easy to enroll at a western university, as you have already studied in an international environment. So, you arrive at an American university and nothing really changes – you are in your element. No frustration, adaptation, stress, and so on. Only your major and topic change, but the culture remains familiar, and that is very important.
Why were you attracted to the New-York University, in particular to the Courant Institute of Mathematical Science?
I was looking for mathematical programmes and universities that also have a reputation of research centres.
The most difficult thing while applying was to explain to the university why I want to change the educational area from economics to mathematics
Of course, one also needs to show excellent grades in mathematical subjects. Some universities, like Stanford, University of Michigan, University of Chicago require applicants to sit an additional exam – GRE Mathematics Subject Test.
When did you have a clear understanding of what you are going to do after university?
When I was in my 1st year on the MSc programme I studied theory and reflected on how I was going to use it. In my 2nd year I understood that this can be applied to Machine Learning, so I started advancing in this area. A few years ago, some applied PhD programmes were opened at NYU. They were created together with the technological companies - Google, Facebook, Uber. In 2017 after 3 years of my work at Google I became one of the first students on such a programme. Currently I am getting ready to acquire a PhD degree. If you are a new employer at one of these companies, they encourage you doing some applied research. You work and at the same time do a joint project with a research supervisor at a university. This project will become your PhD thesis and that will also bring some benefit to the company. That is why within my duties at Google I am engaged in research with my supervisor – Professor Mehryar Mohri, with whom I have been working since my MSc.
Speaking about your work with Professor Mohri, how would you recommend looking for a mentor?
Professor Mehryar Mohri and I have already been working for eight years, and we are doing a great job. Choosing the right supervisor is the most important thing in a research career and as it turned out in my case, in the career in the company. I had great research supervisors at ICEF too – Alexis Belyanin and Sergey Gelman, they taught me many things.
There can’t be universal criteria of how to choose a mentor, except for similarity of interests. From my experience, similar habits and attitudes to work are also very important. If you both like working at night or early in the morning, or solving huge scope of problems at once, or if you both prefer discussing issues while having coffee or a steak at the university canteen, it will double your success. You will spend much time together, and if the rhythms of work do not synchronize, that will lead to conflict and breakdown.
How were you looking for a job and what is the situation on the job market in the engineering area in the US?
Professor Mohri helped me here as well. He works for Google, I applied there on his recommendation to continue research in AutoML. In the technical area, employers are only interested in how you solve their test case. It is not important what kind of person you are, how good your soft skills are – they check hard skills only and if you solve the test, you are in. I showed good results in this respect. Such is the peculiarity of the tech job market. I am sure that in business, economics, and many other areas it is, in contrast, extremely important how you communicate with people.
What is your position?
I took up a job as a Software Engineer, which is a starting position for 90% of tech companies, but the duties are flexible and there are different tasks for this position in each team. I work in the Google Research department. We conduct academic research aimed at their application in internal operational tasks.
Where would you like to grow in your sphere and what are the development scenarios?
As a researcher, I would like to use the theoretical basis that Professor Mohri and I accumulated in Google and during PhD studies to make great ML libraries and solve complicated problems with their help. I can stay in Google, advance career-wise and use our groundwork there. I can do a start-up, and Google will help in this case – the company encourages start-ups and supports its employees in various way.
Have you thought about teaching?
This is interesting but very hard for me. To read material, answer students’ questions – I find it really painful, it is easier to programme and quietly do my job alone with figures.
Is it true that digital people are introverts?
I suppose so.
