Markus, Senior Expert Data Science

About Me

My Technology Skills

blau-neu.pngR + ML Frameworks 

gelb-neu.pngDatabase Design

My Soft Skills

blau-neu.pngTeamwork + Honesty



Technologies I'm proud of working with

computer-chip.png Docker Technology

computer-chip.png Shiny

computer-chip.png Auto-ML-Frameworks


What is your job at Arvato Financial Solutions?

In a nutshell, I can be found at Arvato Financial Solutions in the Data Science department at PAIGO GmbH. My job as a Data Scientist is very complex. The tasks range from classic data analysis to the establishment and integration of the latest machine learning and AI processes. We work closely with our specialist departments, whose needs present us with new challenges almost daily.


How can I imagine this exactly?

I would like to answer this question with a case that we recently worked on. It was about the "digital concern recognition" in our incoming mail area at PAIGO. You have to imagine that we receive countless letters every day, which were previously scanned in using text recognition (OCR) and then assigned manually. Our task in the Data Science team was to develop a system that automatically recognizes the contents of the digital texts and assigns them to a request. Based on these concerns, an automated assignment to the corresponding clerk could then be achieved. For this project we mainly used the programming language "R" in combination with docker technologies. In close cooperation with the departments, we explored, analyzed and structured the available data. Using various NLP (natural language processing) methods and machine learning, we then built the models that were finally integrated into production with the help of docker.

An important aspect of the whole pipeline is also that we use the daily incoming data to monitor our system and continuously improve the algorithms. And this, too, is done in close communication with the specialist departments. This is also how my way of working is structured: a lot of coding, but also a lot of communication and coordination. Because if you don't understand data, you can't use it.

You've been with Arvato Financial Solutions since 2014. Does a routine creep in at some point?

No, absolutely not. Our work is extremely variable and also very creative. Otherwise, I guarantee I wouldn't be with the company for so long if my work wasn't so flexible and varied. The tasks change almost every day! We are currently working on "Speech to text" and improving our chatbot. These are very exciting tasks.


Can you give us some insight into your department?

Of course. We are 7 people in total. My supervisor Nicole and 6 other team members. Within the team we are very connected (even in Corona times). We have daily team routines where everyone says what they are working on. We are usually set up in a way that at least 2 people are working on one topic. Just to challenge, support and validate each other a bit. It's not a job where you always know immediately what to do. Often you have to think in advance "how could I solve this? What methods do I need to use to get ahead of this?"

Furthermore, we also work closely with the company's departments to define projects and goals together. Our department wouldn't work if we were just told from the outside "this needs to be done now" and we bluntly start coding without penetrating the data and the subject matter. Instead, we work very interactively with the departments. And that with a lot of fun and success.

Furthermore, it is exciting to mention that we are very well networked with the other Data Science departments in the Bertelsmann Group. There are various initiatives, such as Data Talks, to bring the different divisions of the Group closer together. This not only broadens our horizons, but also gives us the chance to share our experiences with our colleagues in the group.



What's your favorite game?

That is easy to answer. The game is called "Mechs vs. Minions" and is a board game. The game goes back to League of Legends and is a cooperative campaign game, where you play together with your teammates against a (more or less) artificial intelligence. You can compare the approach a bit with "programming". You have to "program" what your character does on the field by means of cards you have, in order to defend your home base. You have to be able to rely on your teammates and work together against your opponent. It's similar in my job, except we don't talk about opponents, we talk about the challenge we face in our work 😉.

Markus Ruschhaupt
Senior Expert Data Science