Hey, Aleksandra. We’re happy to have you as an interview guest. Could you briefly introduce yourself to our readers?
Hello! I am very glad to talk, thank you for the invitation! Briefly about me – my name is Aleksandra. I am the CEO & Founder of Data Science UA.
I was born in Dnipro city, and I’ve got 2 diplomas: one in tech and one in economics. After I graduated, I moved to Kyiv and began working in the IT industry.
Why did you choose data science and how did you get started in the field?
At some point, I realized there were only a few experts in data science.
So, in 2016 I decided to hold the first data science conference in Kyiv. Suddenly, it turned out to be the right decision: we’ve got a lot of attendees and great feedback. There were only a few networking events in this field in Ukraine, and people wanted to know more. Now, together with the team, we hold such conferences twice a year.
The data science conferences were a steppingstone for your company’s development. But what other areas does Data Science UA specialize in, Aleksandra?
Since 2018 we have four main areas: consulting, conferences, education (corporate trainings, mentoring, courses and workshops) and recruiting.
Talking about data science recruiting, Data Science UA is the first agency of its kind in Ukraine. How did you decide to set it up?
Soon after the first conference I got bombed with requests for data-related workshops and recruiting from both developers and various companies. It was clear that the market lacks this kind of specialized agency. So I decided to give it a try.
Pretty much everything we do in Data Science UA is based on market needs. We get different requests all the time. After we get the same request for a few dozens of times, we try to fulfill it by ourselves. That’s how we started not only the recruiting agency but also the consulting branch.
Is high demand the reason why you chose to add a mentoring program to your services, as well?
We communicate with developers across Ukraine a lot, and we noticed their desire to advance in data science. At first, they studied by themselves using some MOOCs like Coursera, Standford Online courses or Udemy but there is totally not enough practice. Therefore, we started our mentoring program: we find a mentor, usually a high-level expert, who can direct and prompt their mentee.
We also have corporate training programs. Companies come to us because we know almost everyone in the field and can help pump their guys. It’s a similar situation with recruiting.
And the conference is always an incredible chance to call tough speakers and show them that Ukraine is a very capable country in terms of development and data science in particular. It is also a platform for developers where they can hear the latest developments and investigations in the data science world and communicate with more than 500 experts from around the globe in one day.
Aleksandra, you’ve certainly created lots of opportunities for developers and aspiring data science professionals in Ukraine. When it comes to company projects, can you share what industries Data Science UA has collaborated with?
We work with companies across different industries, mostly focusing on retail, pharmaceuticals, fin tech, and agriculture.
We’d also love to hear more about the conferences you hold twice a year. Which are the main themes you usually like to focus on?
We’ve already held 6 Data Science UA conferences. We’re now preparing our 7th Data Science UA Conference which will take place on October 19th in Kyiv, and we’ll be very glad to see you there!
We have three stages of the conference: technical, business and workshops. For the tech stage, we invite the best developers with their cases. Usually, talks on the tech stage are full of developing and engineering nuances. Business managers who have successfully implemented data science in their companies come to talk on the business stream. As we are focused on the practical implementation of data science techniques, we ask speakers to present exact cases with metrics achieved. The strongest practitioners lead two-hour workshops for a small audience of developers and enthusiasts.
Aleksandra, are there any specific requirements the speakers must fulfill in terms of education and profession?
We only have a few requirements. First, the speaker should have proven experience in working with data science. Second, the talk should be new and performed on our stage for the first time. And last, we focus on the practical side rather than academic research or general knowledge talks.
Sounds like a great space to learn practical skills! We’d love to attend one of the Data Science UA conferences in the future.
You also hold educational events for aspiring data scientists once a month. What are the best tips you can give them to land their first job?
One of the most important things is tech education. The vast majority of successful data scientists we know come from technical universities all over Ukraine.
So, be sure to know at least one programming language – Python or R.
We highly recommend taking online courses. There are also plenty of good courses for free. And don’t forget Kaggle competitions – they may help you a lot to feel the “real-life tasks”. As I mentioned earlier, we have a mentoring program, so junior data scientists can get a mentor who will help them maintain their desired career path.
Surely it is necessary to attend conferences: listen to talks about state-of-art technologies and implemented cases and get acquainted with the best data scientists from different countries.
Aleksandra, in your opinion, which are the top 3 common challenges most beginner data scientists must face and overcome in their day to day professional life?
I think these challenges are data quality, expected multi-functionality and the curse of the complex algorithm.
During the studies or Kaggle competitions, beginners mostly get clean data sets while in real life it almost never happens. While everyone is saying that 80% of data scientist’s job is to clean and manipulate the data, there are not enough tasks like this during education time.
A data scientist now is expected to be a Jack of all trades meaning to know math, stats, different ML algos, and domain knowledge, as well as being able to present the result to the business client.
The truth is, it’s pretty much impossible to have everything in one person. Furthermore, it becomes much harder to be an expert in all kinds of ML algorithms as there are so many of them. This makes it very hard for a newcomer to understand the learning path and to focus.
The curse of the complex algorithm is the thing that annoys us a lot. Junior data scientists tend to use the most complex algorithms to show they are experts even though using the simplest one may be enough to achieve the goal. Business does not use LightGBM and Deep Neural Networks to develop, let’s say, a credit scoring model or something like this. Linear regression, logit, and other old and simple algos are often unfairly forgotten by the junior guys.
Thank you for sharing such useful insights, Aleksandra.
You also help companies find experts in different areas of IT. Which are the top qualities most companies are looking for in potential candidates? Do you think there are enough development opportunities in Ukraine for people who are just starting their career in the field?
At the moment, Ukraine is experiencing a growth boom in IT, so there are a lot of opportunities for developers across industries and technologies.
I think Ukrainian specialists have good chances on the international arena: they have a strong technical background, huge work experience (a lot of them started working since the 3rd year of the university, that is, from 18 – 19 years) and impressive efficiency. They can penetrate into different technologies easily, know several programming languages and are fluent in English. Our specialists attend the most influential conferences across the globe, publish research papers and win medals in Kaggle competitions.
So, Ukraine is taking the world of data science by storm! Quite impressive!
Now, let’s talk a little bit about business and data. Today, a growing number of companies realize the need for integrating data science into their business and extracting the maximum benefit from it. Based on your experience, which are the most important steps toward implementing a data-driven business model?
Today’s business models and value chains will soon come under increasing pressure.
Businesses are concerned with the impact new business models will have on their results. Decision makers often don’t fully understand what can be done with the previously collected data or which data should be collected to solve their business tasks.
So, I would say that the first step is education. We teach managers what data science is, how it can be used, what it’s like to be a data scientist. It is important to understand that there are no shortcuts from Excel tables to AI.
The next step is finding a good business case to show the value of data-driven decision making. This could be, for example, churn retention for employees or clients, sales forecast or support tickets classification. The success of solving this case using data science techniques will be a reason to proceed with the data-driven approach on the higher levels.
Last, but not least, is interpretation. One of the key team members is a business translator (also called analytics translator). Among the core functions of this expert is interpreting models results into actionable insights for business users.
Final words to our readers?
As a final note, I would like to say that the data science sphere is an amazing growth zone right now. Data science specialists can predict all people’s movements and actions, and this is exactly what the business needs. The most important thing for specialists is not to be afraid, to study and try new things all the time.
P.S. Everything from the “Black mirror” comes true, look carefully!
Thank you for being here with us, Aleksandra! Your work truly matters for data science development in Ukraine and we wish you lots of successful projects in the future!
Read from the original source: 365 Data Science