Interview with Stelios Gisdakis @ Klarna

Hi there,
Thank you for participating in this blog series!

Tell us your name and what is your current title?

Stelios Gisdakis, Data Scientist/Team Lead @Klarna

What are you working on right now, something exciting you can share?

Working on a new Fraud model for the US market.

How did you get where you are?

Studied Computer Science and then came to Sweden for a M.Sc. at KTH. In 2016 I earned a PhD from the EE department of KTH. Then I joined Klarna as a Data Scientist.

What are the most interesting aspects of your current job?

Definitely creating new predictive models. It is really exciting to see the impact a new model can have on the business, even more so in a company like Klarna - once a model is “pushed to production” it takes live decisions immediately! Surely an exciting (and to some extent stressful) feeling.

Is there anything you would have liked to know about being a Data Scientist before starting a career in this sector?

That having the required education and the technical skills is necessary but not sufficient to make you great data scientist. What is also needed is a solid understanding of the industry you work on leading to an understanding of the important business problems that need to be solved. 

What technologies do you believe will become the next ”big thing”, both in the short term and the long term?

In the short term, I believe that Apache Flink is a technology that can really facilitate the work of Data Engineers and Data Scientist especially if the Flink community develops delivers a Machine Learning library.

In the (hopefully not so) long term, I believe that the graph frameworks of Spark and Flink will (and should) mature. This will really boost the data science community since many of the problems we work on could be approached by means of graph theory and discrete mathematics.

Lastly, do you have any interesting books in our sector to recommend for summer reading?

Actually, I have 2.

First, “Reasoning about Uncertainty” by  Joseph Halpern. This book is a more philosophical and academic text on uncertainty and learning.

Second, I was a bit late in discovering Bayesian Methods for Hackers: Probabilistic Programming and Bayesian Inference by Cameron Davidson-Pilon. I strongly recommend it since it is an easy read with many hands-on examples. It is freely available online (http://camdavidsonpilon.github.io/Probabilistic-Programming-and-Bayesian-Methods-for-Hackers/).

Thank you again Stelios!