Dennis Fast joined DATEXIS team as a research assistant in January 2022 and he is currently working with Paul Grundmann, Alexei Figueroa and Conor Fallon on the development of a large transformer-based pretrained biomedical language model.

He first came into contact with data processing during his education as a physical-technical assistant, where he acquired the foundations of statistical data analysis and an awareness of the importance of the reproducibility and interpretability of results during the large number of experiments in various areas of physics.

He then studied Applied Mathematics with a specialisation in Physical Simulations at BHT (formerly Beuth University of Applied Sciences) in Berlin and in March 2018 he graduated from there with an MSc in Computational Engineering with a specialisation in Acoustic Simulations. During this time, he worked in several research teams at the university to immediately translate the acquired knowledge into real-world applications on the one hand and to deepen the fundamental knowledge on the other hand.

Afterwards, his master thesis he joined the European research projects Shift2Rail FINE1 & FINE2 in the field of rail vehicle acoustics as a computational engineer and later as the technical coordinator of the project, where he continued to gather experience in the generation, post-processing, analysis and visuallisation of data.

The growing awareness of the limitations of conventional analysis and simulation methods, the advanced maturity of machine learning algorithms and interest in neural networks have finally inspired him to pursue the second Master in Data Science at BHT in October 2021. There he found a passion for the Deep Learning algorithms in general and for real-world applications of large language models in particular.

In his free time, he mostly enjoys spending time with his family, hiking or cycling. He also enjoys listening to inspiring podcasts with successful scientists and reading positive science fiction books about the bright future to which he hopes to contribute a small part with his knowledge and experience.

Research Interests:

  • Deep Learning
  • Applications of NLP
  • Physics Simulations


  • Torsten Kohrs, Karl-Richard Kirchner, Dennis Fast, Alberto Vallespin, Joan Sapena, Ainara Guiral Garcia, Otto Martner: Sound propagation and distribution around typical train carbody structures. Euronoise 2018
  • Svenja Hainz, Dennis Fast, Siv Leth, Elodie Vannier, Rüdiger Garburg: Noise Assessment of Railway Innovations, Hands on Sustainable Mobility 2019
  • Karl-Richard Kirchner, Dennis Fast, Torsten Kohrs, Haike Brick: Airborne sound source characterization for railway noise predictions - based on vibration measurements and numerical simulations, Internoise 2019
  • Torsten Kohrs, Karl-Richard Kirchner, Dennis Fast, Haike Brick, Ainara Guiral Garcia: Industrial engineering framework for railway interior noise predictions, IWRN13 2019