Short CV

David Peter Federl, M.Sc., has been a research associate at the Chair of Intelligent Embedded Systems in the Faculty of Computer Science at the University of Duisburg-Essen since April 2025. He completed his Bachelor's degree in Applied Computer Science at UDE, followed by a Master's degree in Cyber Physical Systems, also at UDE. In his Master's thesis, he focused on the evaluation of training methods for hybrid digital-analog AI systems.

He is currently employed on the ZaKI.D project. The project focuses on implementation of artificial intelligence on extremely resource-constrained devices such as sensors or other smart devices. This approach helps to avoid data protection issues while enabling new services to be provided directly within the product or machine. The goal of the project is to make artificial intelligence accessible to companies in the Duisburg  and the surrounding areas. Through small-scale projects and training initiatives, the aim is to facilitate knowledge transfer into these companies.

Research

The focus of his current research is on a thorough examination of the opportunities that arise when implementing compression strategies in FPGA accelerators within the realm of machine learning. This investigation primarily aims at enabling the execution of extensive neural networks on hardware systems that face significant restrictions in terms of resources. This is achieved through the innovative use of dynamic parameter reloading techniques, which hold promise for optimizing performance under constrained conditions.