What made you want to focus on data science as a specialism?
Data science wasn’t something I planned to do from the beginning, it is more something I rolled into. I always enjoyed working on mathematical problems which is why I pursued a master’s in econometrics (i.e. a combination of mathematics and economics). However, I missed the applied part and, when I took a minor in programming, I fell in love with coding. I enjoyed this so much that I taught some courses in coding and continued with a master’s in software engineering.
I also started working as a software engineer (on data-related problems). At some point in time, when I had been programming for three days in a row, I realised I was missing modelling and the connection with the business.
For me, data science provides the perfect mix of programming, contact with the business (as you need to explain the results), and mathematics. Working with data can be frustrating at times, because your bugs may not only come from the code you wrote and the model you made, but also from the data itself. However, this also makes the work dynamic and challenging in my opinion.
How does your data science work impact Adyen as a business?
I’m working on payment optimization, which might sound boring, but is actually a lot of fun given the huge number of payments Adyen processes every day. With our machine learning model, we can significantly increase the success rate of payments, which is helpful for our end customers (how annoying is it if you have just selected all your groceries but your card is not working?) and brings value to our merchants, such as Uber, Hello Fresh, and Booking, all of which highlight the importance of career acceleration: the power of mentorship and sponsorship in driving innovation.
Do you have any predictions for how you think the field will evolve in the future?
I happen to have written an article about this topic a few years ago and I still believe this is largely true. I’ve seen the modelling part becoming more automated by an increasing number of Python packages that are available at different stages of model development. I expect that for larger companies which can afford to hire in-house data scientists, the job will remain technical and we can deliver most value by understanding the problem, interpreting the results, and working on projects that are hard to automate. For smaller companies, I expect an increased use of out-of-the-box (cloud) solutions.
Are there any tools, resources, organisations that you find useful for technical development?
The nice thing about data science is that there are many resources (freely) available. I personally like online courses. I can recommend the Stanford course on Coursera (https://www.coursera.org/learn/machine-learning). To gain experience it’s also helpful to participate in some Kaggle competitions (https://www.kaggle.com/) or attend a meet-up to meet other people working in the field.
How do you think data scientists can best prepare for technical interviews?
I think this depends on the skill(s) you think you might need to improve. If that is programming: start a pet project or course about something you like to do. For modelling I would recommend practicing with some Kaggle competitions. For pure data science knowledge I personally like reading books. My overall recommendation would be that you practice in such a way that makes it engaging for you, which really varies from person to person. So try some various methods and find out what works best for you.
