ok.tech: Data Talk # 2

On August 7, ok.tech: Data Talk # 2 will take place at the Odnoklassniki Moscow office, this time the event will be devoted to education in Data Science. It is not a secret for anyone that now such a hyip is around work with this, that only the lazy did not think about getting education in the field of datasaens. Someone thinks that without a university education to become a datasaientist is impossible, there are supporters of the opinion that one can learn to work with data using courses, others take the position that a good datasaentist is a practice and a many-sided approach. We will gather representatives of different opinions on our site and give them the opportunity to discuss this topic. The event will be held in the format of a discussion between the speakers, this time with us will be Evgeny Sokolov (HSE, Yandex.Dzen), Dmitry Bugaychenko (OK.ru), Peter Ermakov (Lamoda, DataGym ), Dmitry Korobchenko (Nvidia, GeekBrains, SkillBox, Digital October) and Victor Kantor (Mail.ru Group, Data Mining in Action ). We invite everyone who is interested in the topic of education in DataScience to join the event and express their point of view. We studied at the courses, come and tell us what it has given you, think that it’s impossible to analyze data in the analysis, come and tell you why you think that a datascenist should be able to write in the prod, come and discuss.

→ Registration for an event
')
Under the cat expert opinion and schedule.

Evgeny Sokolov, Higher School of Economics, Yandex. Dzen

Now there are many options for teaching data analysis - there is something closer to the “technical school”, where they are taught to simply use ready-made tools, if there are opposing opinions that you should look at ML as a mathematical thing, and not as a craft. I think that first of all you need to learn the craft, because without this it is impossible to motivate a student, and even then at work he will use these skills 80% of the time. But at the same time, it is extremely important then to teach him the correct thinking and deep understanding of the methods - without this, the student simply will not become competitive in the labor market.

Dima Bugaychenko, OK.ru

According to DS Education, I would single out several important “Challenges” that distinguish DS from other areas. Firstly, this is a dynamic - everything is changing very quickly and therefore one cannot learn, get a diploma and become a DS, one can only constantly learn to remain DS. Secondly, it is a synergy of very different disciplines - you need to understand the mathematical essence of the methods, and be on technology with you (if we are talking about DS and not about a monkey poking at the XGBoost stick). And thirdly, this is a very high demand for educated DS from industry, coupled with a large gap in expectations between the industry and the academy in Russia, which, among other things, leads to the emergence of a large number of “schools” from large market players.

Peter Ermakov, Lamoda, DataGym

I love teaching very much, I love the moment when I can tell the difficult things in simple language, and see the understanding in my eyes. Over the past 10 years, I managed to reproduce in 26 launches of three commercial courses, two universities, within the company and conduct an open educational project. And now I am engaged in the creation of a commercial 3-month course on machine learning DataGym.ru. All types of education are good in their own way. And commercial courses are no exception. These are other opportunities, another threshold of entry, another level of motivation and time-consuming.

Dmitry Korobchenko Nvidia, GeekBrains, SkillBox, Digital October

My position is that there are no directions that would have all the possible advantages. I can not say that one thing steers, but this is another - no. I am more for matan, linal and normal math. I don’t really like it when people use tools without understanding how they work (at least on an average level). But I think that in some business cases this will be justified. Especially considering the democratization of I.I. Regarding Kaggl, I can say that I know a lot of people (and myself from such) who have developed quite well in the region without resorting to this resource at all. But I think that he still gives an additional boost in certain skills.

Victor Kantor, Mail.ru Group, Data Mining in Action

Each year, as part of the offline course Data Mining in Action, about a thousand people get acquainted with machine learning. In the online courses launched only by us and our colleagues (and there are obviously many other courses in the world), over the past three years about 100 thousand people have participated. Of course, those who do not just “get to know”, but reach the end and become, for example, Junior Data Scientist, are significantly less, but still a crazy number of people are coming to data analysis right now, so hiring a person for the initial position very difficult. But from the middle level and above the problems start - the search for an employee immediately becomes long, painful and, as a result, expensive. What to do with this is the question that I deal with now.

schedule

18:30 - 19:00 - Registration of participants
19:00 - 19:05 - Intro from Alexey Chernobrovov
19:05 - 20:00 - Controversy on education in Data Science
20:00 - 20:20 - Coffee break
20:20 - 21:30 - Continuing controversy

→ Registration for the event