In cooperation of the Harz University of Applied Sciences and PITERION, our colleague Katharina Herrde has successfully completed her Bachelor in Media Informatics (Department of Automation & Computer Sciences). During her studies, Katharina focused on deep learning and artificial neural networks around the topic of Natural Language Processing (NLP).
Commenting on her bachelor thesis, she says: "My work addresses a contemporary and relevant topic, but not many companies deal with artificial intelligence or machine learning. That's why I was glad that PITERION gave me the opportunity to work in this field."
Katharina's research was based on the motto HR meets Machine Learning: Although training and education is beneficial for both employees and employers, some employees do not use the opportunity. One possible reason for this is the time it takes for employees to find and select suitable training. Another problem are the resulting costs for the company.
As a potential solution, Katharina has designed and developed a "Recommender System" for trainings as part of an internal career portal. With the help of this system, each employee is suggested individually suitable development actions from a training catalogue. Factors such as costs, availability, competencies, project assignments and the employee's current and desired position are taken into account. The recommendations are also based on the experience and knowledge of other colleagues.
Collaborative, model-based recommendation systems offer the advantage of being able to recognize and use connections between user groups and their preferences. This process is time-consuming for large amounts of data, and if only little information is available for a user or a training course, the quality of the suggestions suffers.
For this reason, the recommendation system of the internal career portal works hybrid. Algorithms for processing natural language and content-based suggestion generation are used to extend the collaborative approach and to quickly generate suitable recommendations using both large and small data sets.
"Thomas is always open to all innovative ideas and could give me new impulses if I got stuck on a problem. My other colleagues at PITERION have always given me their backing. Whenever I had problems with my development environment or questions about the structure of our career portal, I could always turn to them. That doesn't just apply to the developers, but also to the others. For example, it was good to hear how my undergraduate colleague Tobias from Sales Management made progress in his own thesis. This allowed us to motivate each other - and now we're both full-time employees at PITERION".
Katharina presented the great results of her work in March at the CC Partner Symposium at the Munich University of Applied Sciences. Since April she has been an integral part of our team in the field of "Industrial Software Development" in Wolfsburg.