Earlier this year, like most undergraduate students before holidays, I was wondering what I should focus on in regards to my career development. I decided to try data science for a few reasons – it is not only one of the hottest, best-paid professions right now, but also allows the development of two areas I am interested in: programming and business.
As a 2nd year undergraduate, I was lacking the necessary experience to get a real job as a data scientist. I knew, however, that developing these skills by taking online courses or university classes would take much longer than I’d like it to. Despite studying computer science at one of the top universities in Central and Eastern Europe, I convinced myself that taking part in a data science internship is the fastest and the most promising way of learning skills essential for a data scientist.
I have explored plenty of options trying to find a company that would hire me as an intern, finally choosing Semantive. They accepted my lack of data science experience and valued other skills, such as web development, team management and attention to detail as far as architecture design or code refactoring are concerned. Initially, I was placed in a team of 4 interns, where only one of us had a real Data Science background. Others, just like me, were recruited for other relevant skills they possess. We completed a few projects together, including 2 Kaggle competitions and internal serverless applications, learning a lot with the help of experienced mentors.
After such experience, I am confident that finding appropriate internships is a way to go when it comes to developing skills, especially the ones your university fails to provide. I wanted to share some reasons for taking such an approach, and I believe they can come in handy if you’re deciding what internships to apply.
Working on relevant projects
Learning by doing is undoubtedly the most effective. I enjoyed many of the university projects that I worked on, but after having completed one or two of those, I realized that I needed to do tasks that are actually relevant for someone else. Don’t get me wrong – I think it is important to continue studying, but part-time work or an internship can be equally challenging and rewarding. The fact that your efforts are not in vain is an excellent motivation for me, and without knowing what is important in the industry I would have never known what to focus on at the university.
Keeping up with the latest research
Beginning to learn a new field is always challenging – lots of buzzwords, misinformation or outdated courses are ´the noise´ you have to get through. Keeping up with all of the latest research was impossible for me, and I quickly realized how priceless it is to know people that are familiar with state-of-the-art solutions. Only work and scientific research offer such an access to information. If you’re not a Ph.D., finding an internship is the only option. Once you encounter the right company, projects can be well prepared and technologically as interesting as the research itself. The Access to AWS or another cloud, reusable tools, and plugins for common tasks as well as a knowledge base and the access to professional mentors are some of the factors that made my work at Semantive much more convenient.
Understanding models usage in the real world
Most machine learning courses focus on the scientific part of data science – which is crucial, but there are also other technical challenges that people face, such as inference speed or model interpretability. Without the internship, I would not have realized how important these factors are in applied machine learning. Integrating model with applications or processing data to correct format at scale is a job for someone like a data engineer, not a data scientists. You never know what you’ll enjoy more. I was lucky enough to work on some of these problems as part of my data science internship, which equipped me with a significantly wider range of experience than I could get at the university.
Learning from experienced professionals
I am lucky to be studying at the University of Warsaw, which arguably has the best computer science course in Poland. I always take every opportunity to learn from the best, but there are occasions when I just don’t have the right support. I can only imagine how bad it can get in some of the other schools. During an internship, however, you are always in a smaller team than a class, and if a company hires interns they logically have more experienced people too – otherwise, it would not make any sense for the company itself. At Semantive, we got a different mentor for each of the projects, for the Kaggle competitions we had the most senior data scientists in the company. They would drop in on our daily standups and answer all our questions, which made a huge positive impact on my development as a data scientist.
Exploring various career paths
Since data science is a really fresh field, you will work with people from multiple backgrounds and focus on a variety of problems. You will also find value in skills other than model design. Some of my colleagues liked to focus more on the data engineering side whereas others preferred to stick to the parts of data science. For me, improving experiment efficiency, managing a small team of interns and presenting our results as well as writing blog posts were the most enjoyable moments.
Earning more money
Data science is a position that requires an unprecedented combination of skills, and not surprisingly pays even more than software development roles – according to Glassdoor, it pays about a half more. While at junior or at intern levels these differences are not so significant (Semantive pays the same to data science and full stack interns), it is relatively easier to get an internship in data science than in more traditional areas with a stable demand of skilled students. As I mentioned, I got internship offers despite my lack of data science skills, simply demonstrating the willingness to learn. Also very helpful is the fact that I have other skills potentially useful in a data science team such as programming in Python, following OOP principles and best practices, leadership and project management. All of these allowed me to develop the data science skills I wanted, and a similar approach to internships will enable you to develop yours. The only important thing is to have it clear what new skills you are willing to learn and what other useful skills you can bring to the team.
At Semantive, we are looking for people eager to start or develop their careers in data science. We offer internship positions that do not require professional experience in data science as well as junior and senior roles for people with proven skills in data science and machine learning. Visit our careers website to learn more or apply.
Written by: Krzysztof Kowalczyk, Data Science Intern at Semantive.