How this physicist became a data scientist: Story of Kaggle Grandmaster Laura Fink

For me, coding is a creative process like art.

Published on May 16, 2022

by Sri Krishna

Kaggle Grandmaster Laura Fink is the head of data science at Mircomata. For this physicist turned data scientist, hackathons are all about having fun. “As long as there are teammates to collaborate with, I always like hackathons,” she said.

In an exclusive interview with Analytics India Magazine, the physicist turned data scientist spoke about her Kaggle Grandmaster journey.

AIM: How did your fascination with algorithms begin?

Laura Fink: I was not interested in coding right from the start but rather in natural sciences and math. For as long as I remember, I wanted to understand the world and the differences and similarities between living and inanimate objects. For this reason, I studied physics.

For my master’s thesis, I optimised a program to remove background noise from video data of fluorescent cells. This was my first contact with code and machine learning. And I loved it! For me, coding is a creative process like art. It’s real craftsmanship.

AIM: What were the initial challenges, and how did you overcome them?

Laura Fink: Well, I had no prior knowledge of coding when I read the code of my teammate who wrote the program. I had no clue. It was like reading a book in a foreign language. So I bought a small book about coding in Java to solve this problem. After that, I just jumped into the code and started implementing my ideas.

AIM: What about coding excites you the most?

Laura Fink: You create something that processes information and does something with it. It’s like an awesome machine that you can copy as much as you want and that can easily be changed and shared across the community. I especially like the idea of open source.

I don’t have a special ritual to start coding, but I enjoy it the most in a relaxing atmosphere.

AIM: What does your machine learning tool stack look like?

Laura Fink: Usually, I start to explore the data using matplotlib, pandas, NumPy and seaborn. After that, it depends on the problem. For computer vision, I like both TensorFlow and PyTorch. Otherwise, I often use scikit-learn. I also like to explore other tools, and Kaggle is an awesome sandbox to try out new packages and get your hands dirty with new tools.

AIM: How do you prepare for your first hackathon?

Laura Fink: If you have no prior coding experience, it might be good to do a small tutorial first or read a book to help you get started. This way, you can focus better on implementing your ideas during the hackathon instead of struggling with the basics.

AIM: What’s the worst experience you ever had as a coder?

Laura Fink: The toughest challenge for me when I was a beginner was understanding git. The worst experiences or moments that made me sweat were related to code merging with git.

AIM: What drew you towards Kaggle? How has your journey been so far?

Laura Fink: One day, a colleague told me about Kaggle, and I was immediately hooked. At that time, there were only competitions, and notebooks were only published to share ideas–and not to get upvotes. It felt like a game that I was always looking for. Since then, the platform has evolved a lot, and over the years, I have recognised that Kaggle is good for having fun and improving my data science and machine learning skills. To learn something new, it’s important to fail and adapt. If you are always in your comfort zone and there is nobody to challenge, you will hardly improve yourself.

AIM: What was your first Kaggle competition like?

Laura Fink: I felt overconfident and thought it would be easy. But after making my first submissions, the truth hit me. Courses at the university were good to get started, but I needed practical experience to improve myself. I like that it is not easy to climb up the leaderboard. But, on the other hand, it is a lot of fun to work on myself to become a better data scientist.

AIM: How did it feel when you became a Kaggle Grandmaster?

Laura Fink: My Grandmaster tier is related to the notebooks branch of Kaggle. I used notebooks to improve myself and learn new skills and tools at my own pace. By sharing my learning experiences with the community, I connected with other passionate Kagglers and was able to shed my fear of showing my ideas. With job and kids, I often worked on my notebooks in the evening/night when I had some spare time. When I became a Grandmaster, I felt very happy and sad. It’s the same feeling you get on defeating the final enemy in a video game. I’m very glad that there are still branches left that I can explore to collect more tiers.