Upcoming Talk: Dr. Jonas Rende and Thomas Stadelmann of DATEV eG on NLP@DATEV: Setting up a domain specific language model

Symbolic picture for the article. The link opens the image in a large view.

We will be hosting an industry talk on “NLP@DATEV: Setting up a domain specific language model” in Software-Anwendungen mit Künstlicher Intelligenz (SAKI). The talk is free and open to the public.

  • by: Dr. Jonas Rende and Thomas Stadelmann, DATEV eG
  • about: NLP@DATEV: Setting up a domain specific language model
  • on: July 14th, 2021, 16:15-17:45 Uhr
  • on: Zoom (link after registration)
  • as part of: SAKI

Abstract: In their session, Jonas and Thomas talk about the first DATEV domain specific language model. Their talk cover a short introduction to NLP, the success of deep-learning in the field of NLP, the difference between word embeddings and contextualized embeddings, the importance of transfer learning, the benefits of domain specific knowledge and potential use cases of their language model. Moreover, they explain how they evaluate their proof of concept is successful. Towards the end of their talk they dive-deep into the BERT model.


Jonas Rende is a senior data scientist at DATEV eG working in the customer centric design department. He uses machine learning methods to extract customer needs out of vast text and user behavior data. Together with Thomas Stadelmann, he is working on automatically generating insights out of customer feedback. In addition, he is laying the foundations for a customer experience platform. Before joining DATEV e.G. Jonas was a research and teaching assistant at the department of statistics and econometrics at the University of Erlangen-Nürnberg. Jonas holds a master’s degree in economics and a PhD in statistics from the University of Erlangen-Nürnberg.

Thomas Stadelmann is a senior data scientist at DATEV eG., where his work focuses on Information Retrieval, Neural Search, Query Log Analysis and A/B-Testing. 10 years ago Thomas started as a software engineer at DATEV and over the time his passion for data science continuously increased. His current research interests include Natural Language Processing, Machine Learning with focus on Deep Learning, Search Evaluation, Datastructures, Systems Reengineering and anything computable. Following his Bachelor’s degree in the dual system, Thomas got his Master’s degree with distinction from the Otto-Friedrich University of Bamberg. He was awarded the Best Short Paper Award of the BTW 2015 conference.