Dr. Julian Merten, age 36 and father of a small daughter, holds a doctorate in theoretical astrophysics and has taken over the "Data Science" team for Finance & Controlling at Mercedes-Benz Cars recently. He tells us what astrophysics has to do with his new job, and how the Daimler brand is seen in the IT world.
Hello, please introduce yourself briefly to our readers.
Hi, my name is Julian Merten, I am 36 years old and have a doctorate in physics, or astrophysics to be exact. Or even more precisely, I'm a cosmologist. I studied and obtained my doctorate in Heidelberg, where I also spent a short post-doc time. After that I went to the Caltech [California Institute of Technology] in Pasadena near Los Angeles for four years, where I also worked on the Jet Propulsion Laboratory for NASA. I then had teaching and research assignments in Oxford and Bologna. Now, I just recently started with Daimler.
Sounds like you're a real scientist. What exactly does cosmology research into?
We look at the cosmos as a whole, so not individual planets, stars or galaxies but the entire Universe. We know a few things about it, but there's a great deal we don't know. We know how old it is, and also that there was once a state where the cosmos was absolutely tiny, energy-charged and hot. But for example, we don't understand why the Universe has expanded faster than previously for roughly one billion years. And as we don't know why this happens, we refer to the cause as "dark energy".
Daimler is undoubtedly a strong brand and a respected name in the IT world
So to what use can you put this knowledge at Daimler, except perhaps to worry your colleagues unnecessarily?
[laughs] Yes, naturally you immediately ask yourself what an astrophysicist is doing at Daimler. I am head of the "Data Science” team in the department “Data & Analytics” for Finance & Controlling at Mercedes-Benz. This is about data-based and analytical models for better prediction of financial processes, so as to actively support our controllers in their work. One important example might be estimating so-called "vehicle residual values" in different countries. Naturally such predictions have also been carried out by colleagues at Daimler in the past. But in addition we offer the input from a machine learning model: we teach computers to reveal interconnections that humans can't recognize. The combination of human experience and intuition, supported by the computing power of a machine model, produces optimum results.
And astrophysics helps you with this?
Not necessarily the content, but the methodology does! In astrophysics too, more and more data are collated and used, which means that the numerical models that describe these data are increasingly complex. And that takes us directly to machine learning: my research involved a great deal of computer science and high-performance computing. In fact I wrote machine learning codes and algorithms on a daily basis – but in a cosmological context. I'm still the nerd who says "Hey, machine learning is fun!", and now I enthusiastically apply my knowledge and models to financial processes.
Did you have to undergo further training for your job here?
Naturally further training is important, so I continuously expand my knowledge of finance and controlling. That's the only way to understand the new context, interpret things correctly and do a good job. The most important thing is this though: I'm new here, but not alone! I work together with a wonderful team that helps me to understand the various aspects and how they interact. For my part I contribute my expertise in machine learning. As a whole we are a superb product, I would say.
Why did you choose to work at Daimler?
I actively applied to join Daimler. I went to the Careers page and entered the search term "machine learning". That's how I became aware of this position. I wanted to return to Germany (so did my very adventurous American wife), so it made sense to apply for this job. After all, Daimler is a very strong brand, and also has a good name in the IT sector. Daimler stands for quality and first-class specialists. The group is so large that not everything is all about vehicles - there are many other interesting products and application areas.
The company is large enough to also offer a plethora of other interesting products and use cases.
Please tell us a little about your team and its work.
My team consists of trained data scientists and mathematicians, i.e. people who are familiar with machine learning and statistical models, and colleagues from the financial sector. This varied mixture makes the difference. Our main area is so-called data modeling; we write the algorithms for the "machine", which has a precise look at the data, models them and finally also interprets them.
What profiles are you looking for in your department?
We are particularly interested in IT people with a profound understanding of machine learning, and "deep learning" is also welcome.
Is it important to be proficient in German?
Not at all! We are an international team, and I worked in English-speaking countries for ten years myself. We have a rule in the department: if not everybody in the room speaks German, we switch to English.
Enough about business: It's 8.30 p.m. on Saturday.Where are you usually to be found?
I'll probably be slumped on the sofa with a cold beer in my hand. Because it can be tiring to get a 16-month old child into bed!