Seb***Dr.

Deep Neural Networks (DNN), Embedded Systems, C, Software Development

Director AI Competence Center at NXP Semiconductors

• Location: Munich, Bavaria, Germany
• Duration: 2023-07 to Present
• Description: Building and leading the NXP AI Competence Center and its high-performing research and engineering team at the central NXP Chief Technology Office.

Research Engineer for embedded AI at NXP Semiconductors

• Location: Eindhoven, North Brabant, Netherlands
• Duration: 2021-02 to 2023-06
• Description: Scouting, evaluating and developing embedded AI technology in the Automotive System Innovations team at the Central Technology Office of NXP. My current research focuses on hardware-aware neural architecture search targeting resource constrained embedded systems. Main evaluation use-cases lie in the computer vision domain.

Research Engineer for embedded AI at Bosch

• Location: Stuttgart Area, Germany
• Duration: 2019-03 to 2021-02
• Description: My research work focused on: * Methods for embedding deep neural networks (DNNs) in resource constrained systems. - Techniques: Network compression by quantization, distillation, pruning, neural architecture search - Applications: Automotive computer vision (SemSeg, ObjDet), activity recognition, person identification * Number representations for DNN algorithms. * Hardware architectures for efficient processing of DNNs in embedded systems. - Evaluation on FPGA-based prototypes As a project-lead from January 2021, I formed, motivated and guided an enthusiastic team of highly-skilled research engineers and students to design a scalable, efficient and robust hardware-accelerator for DNNs. Together, we achieved the major milestone of transferring our development to internal clients in the Bosch business line. Together with the business line the development will eventually be made part of a distinguishing hardware-IP portfolio of Bosch.

Ph.D. Candidate at RWTH Aachen University

• Location:
• Duration: 2016-05 to 2020-01
• Description: * Ph.D. supervisor was Prof. Dr.-Ing Gerd Ascheid of the Institute for Communication Technologies and Embedded Systems. * I conducted my research at corporate research of Robert Bosch GmbH.

Ph.D. Candidate with focus on efficient embedded AI at Bosch

• Location: Renningen, Stuttgart Area, Germany
• Duration: 2016-05 to 2019-02
• Description: * I conducted my research in the field of efficient hardware acceleration of deep neural networks (DNNs) for vision-based advanced driver assistance systems (ADAS) at corporate research of the Robert Bosch GmbH. * Specifically, I developed novel quantization techniques to efficiently deploy DNNs on dedicated accelerators. Moreover, I proposed novel number representations and analyzed them with respect to their suitability for implementing DNNs. * Furthermore, I designed and implemented a hardware accelerator for large-scale DNNs on an FPGA. This accelerator was equipped with efficient processing elements exploiting the proposed novel number representations.