As the Data Science Bootcamp expands, we onboard new team members. But becoming part of this team is not an easy task given the high criteria and standards imposed by the existing members.
That is why today we are pleased to introduce Igor Trpevski. He is an instructor of the Machine Learning and Python modules at the Bootcamp. Igor is a data science consultant with many years of experience in the industry, four of which he has spent in the renowned company Procter&Gamble.
On March 23rd, Igor will hold a free online lecture, “How to achieve a successful Data Science result on Kaggle”, through which he will convey specific techniques and practices used at the “Data Science Olympiad”.
Brainster: Igor, your professional portfolio contains seven years, as a researcher at MASA and four years as a Data Science consultant, at Procter&Gamble. What are the differences when it comes to Data Science for research and in business/industry?
Igor: At the Academy, the problems you can work on may or may not have direct practical value. Instead, there must be an existing product in the industry, that value the end consumer. It is rare for companies to fund research purely for the good of society (for example DeepMind). Also, there is much more pressure on deadlines.🙂
Brainster: On March 23rd, you will hold a webinar on the topic “How to achieve a successful Data Science result on Kaggle?”. What can viewers expect from this webinar? Why did you choose Kaggle as its theme?
Igor: The viewers will learn the most commonly used steps from Data Science that they should take in a Kaggle competition, to achieve a high score. So data preprocessing, feature engineering, cross-validation, hyperparameter tuning and combining different models (ensembling).
All this will be demonstrated with a concrete example from a previous Kaggle competition.
Brainster: You hold the Python and Machine Learning modules at the Bootcamp. What are your favourite Python IDE and a favourite Machine Learning library?
Igor: Lately, I use JupyterLab more and more. As for my favourite library, it is hard to choose because there are many good ones. But if I have to choose, then it will be Fast API.
Brainster: Why do you think Python is established as the industry standard for Data Science? R is also in competition, but the statistics say that Python is more used. What is the reason for that?
Igor: Probably this is because many people moved to Data Science, from the Computer Science programmes, where Python is learning more often.
R is traditionally teaching in Statistics programs that have produced smaller staff. From a technical point of view, there is currently no reason to choose Python over R. I think it is easier for the industry to find Python staff.
Brainster: From your experience so far, how satisfied are you with the students progress on your modules?
Igor: There is a difference between student to student, but I can say that I am very pleased with those who are active.
Admissions are open for the Autumn Data Science Brainster Bootcamp. Join us and future-proof your career by building a top-notch and job-ready portfolio.
Brainster: Machine Learning has become a key driver behind the accelerating digital transformation. In your opinion, are we making the most of this potential? What automation degree can we expect in the next ten years?
Igor: There are dark and light points from the application of Machine Learning in various automation segments.
I will point out the exceptional success of Deep Learning algorithms in medical diagnostics development.
An interesting example of this is the algorithms that can diagnose heart attacks only by the sounds that a person makes while sleeping and are recorded on its phone app. Imagine that a patient with irregular snoring in a dream (as a heart attack sign) results in an automatic ambulance call to the patient’s home.
On the other hand, the algorithms used in today’s social media content referral systems have frightening but often overlooked consequences. This social polarization is on a scale that we have never seen before, at least not since the end of World War II. These algorithms reinforce prejudice among large masses of people by offering them content full of fake news.
Brainster: Data Science as a multidisciplinary field requires a lot of learning and dedication. We already know this from our students. But we have never asked the instructors about their career journey. What can you tell us about yourself regarding this? How much dedication and work did it take for you to master these skills?
Igor: Honestly, there is no shortcut. You need to dedicate time to Data Science. I would say that I was lucky to have a good mentor with a lot of patience while working at MASA, which enabled me, and several other colleagues to gain extensive knowledge in the years spent at MASA.
Brainster: Do you think the fear of the general public is justified about Data Science and programming? What is the reason for that? As a professional, would you recommend it to all those interested in a career in this field?
Igor: I do not think it is justified at all. People should try to learn some basic things without any pressure, and if they are interested, go ahead and spend more time on the details.
There are many examples of people who transition to Data Science from different backgrounds. I would suggest anyone who is afraid to read a bit on Wikipedia about Jeremy Howard’s career. The man studied philosophy and has started his career in management consulting. About 15 years later, he won several Kaggle Challenges. Now he holds Deep Learning courses for people who have never programmed in their lives.
Brainster: What do you do in your spare time?
Igor: I have a big flaw, the fact I have little freedom to organize myself. Generally, my wife and I both trying to work for the same hours, so we can spend our free time together. During this period, I try to spend the evening reading the Witcher books. It is much easier to fall asleep when a person moves into a fictional universe. We spend enough time thinking about real problems.
Brainster: Finally, what are your message to current and future students at the Data Science Bootcamp? Maybe this webinar would be a great starting point for beginners?
Igor: You should devote hours and hours looking at code from existing algorithms and trying to understand it in detail. Then commit to one Kaggle challenge and try to reach a result that will rank in the top 30%. That’s great for a start.
If you have ever thought about a future-proof career in Data Science, have a look at our remote Bootcamps.