The increasing adoption of Software-Defined Vehicles (SDVs) is transforming automotive software development, demanding greater flexibility, scalability and continuous evolution of software systems. However, current development practices remain constrained by legacy architectures and processes. In this paper, we identify key challenges that hinder this transition, including platform dependency, tightly coupled application software, premature optimization driven by early design decisions, toolchain and vendor lock-in and late-stage resource consumption analysis. These limitations reduce portability, increase system complexity, and delay feedback cycles, ultimately impacting development efficiency and innovation.

In this PhD work, we are systematically investigating these challenges through real-world examples drawn from a large automotive group. This empirical perspective enables a deeper understanding of current limitations and supports the development of more modular, flexible, and analysis-aware approaches for automotive software systems.

Muhammet USLU is an industrial PhD candidate at CARIAD SE in Germany, the software division of the Volkswagen Group, where he focuses on next-generation automotive software development for brands such as Porsche, Audi and Volkswagen. Born in Türkiye, he has built an international academic and professional career across Europe.

He has gained hands-on experience in applied machine learning and embedded systems through roles at leading research and industry institutions. At Bosch Research in Stuttgart, he worked on estimating the energy efficiency of deep learning algorithms on hardware accelerators and optimizing their performance using advanced compiler techniques. Prior to this, he contributed as a research assistant at Fraunhofer IIS, focusing on deep learning, neuromorphic computing and quantization-aware training.

Earlier in his career, he completed internships at Valeo Siemens eAutomotive in Erlangen Germany and Aptiv in Krakow Poland, where he worked on hardware development for electric vehicle systems and system engineering processes.

He holds a double Master’s degree in Computational Science and Engineering from Friedrich-Alexander-Universität Erlangen-Nürnberg in Germany and Università della Svizzera italiana in Switzerland. He previously earned his Bachelor’s degree in Electrical and Electronics Engineering from Çukurova University in Türkiye.