P7S1 x Munich Datageeks
Together with the non-profit organization Munich Datageeks, we hosted a Meetup that focused on the topics of Data and Ai at our campus in Unterföhring.
The idea behind the Datageeks community and their Meetups is: to bring together people who want to make an impact in the digital world and to create a knowledge transfer. For us, this was a good chance to exchange ideas, learn from each other, and give insights into our P7S1 Tech & Data cosmos. Everyone was welcome — from students and academic researchers to software engineers, statisticians, data scientists, and entrepreneurs.
After a short intro session by the Munich Datageeks and our colleague Richard Seitz, SVP Data Products, the evening kicked off with a lot of insights into the world of Data. The main focus of the evening was on the topic of “machine learning” and its many possible ways of application.
The world of machine learning at P7S1
Our colleagues Atef Attia (Machine Learning Engineer), Malte Blumhoff (Cloud Engineer), and Manuel Heller (Data Scientist) gave a first insight into the exciting topic during their presentation. How can ML be used for our TV advertising? Among other things, machine learning can be used for our user tracking data to develop a predictive model that estimates the potential website visits generated by future TV spots. In addition to the predictive model, our three colleagues also presented the creation of a state-of-the-art pipeline to facilitate large-scale model evaluation. At the end, they gave an overview of the AWS infrastructure that powers AI-driven products developed at ProSiebenSat.1.
“By using predictive models, we help our clients to optimize their TV advertising campaigns. We can predict the best TV spots to book — this way advertising and budget can be used more effective.”
From TV advertising to biology: the different possibilities of ML
Another area where machine learning is being used was presented by guest speaker Data Scientist and Researcher Furkan M. Torun. He demonstrated how machine learning is also an important tool in healthcare for analyzing complex and large biological data sets. A common problem: the lack of reproducibility and transparency.
In his talk, Furkan explained how this can be addressed with the help of “OmicLearn,” a web-based open-source platform for machine learning.
After the presentations and the interesting input, it was all about networking. At a get-together, the guests had plenty of time to exchange ideas and make new contacts over drinks and food. We would like to thank the Munich Datageeks and all guests for this insightful event and the mutual exchange!
