PhD Student AI-based Automated Driving (f/m/d)

Work Environment

As part of an interdisciplinary research project, we are investigating the potential of Visual Language Models (VLMs) for use in automated driving systems. While VLMs already show impressive skills in semantic scene interpretation, they have so far lacked a deep understanding of driving dynamics and the physically realistic control of vehicles in road traffic. The aim of the doctorate is to expand a VLM-based driving system with knowledge of driving dynamics. The first step is to analyse the areas in which current models have deficits. Based on this, methods are to be developed to close these gaps - for example through targeted training data, hybrid model architectures or the integration of physical models. Do you want to help shape the future of automated driving and take AI systems to the next level? Then apply now and become part of our dedicated research team!

Possible Tasks Within This Role

• Analyzing the driving dynamics capabilities of existing VLM-based driving systems
• Developing benchmarks for evaluating physically realistic vehicle control
• Integrating vehicle dynamics models into multimodal AI systems
• Fine-tuning large language models for domain-specific applications
• Evaluating through simulation and transferring into real-world driving scenarios

The final doctoral topic is defined together with the professor. The prerequisite for the cooperation is the confirmation of supervision as well as a confirmed delimitation of topics by the professor of a university or research institution entitled to confer a doctorate.

Qualification Requirements

• Good to very good university degree (Master's or Diploma) qualifying for doctoral studies in computer science, robotics, mechanical engineering, electrical engineering, cognitive science or a related field
• Sound knowledge in at least one of the following areas: Machine learning, computer vision, automotive engineering, control engineering
• Experience with deep learning frameworks (e.g. PyTorch, TensorFlow)
• Good knowledge of programming languages (Python)
• Interest in interdisciplinary Research and complex technical systems
• English language level B1

Keywords

Dissertation, Generative AI, End-to-End AI, Foundation Models 

What We Offer

• Attractive salary & 30 vacation days (+ 24.12. and 31.12. off)
• 35-hour week, flexible working hours, remote work
• Special conditions for the purchase and leasing of vehicles
• Free seminars on scientific work and interdisciplinary qualifications
• Participation in the doctoral network for scientific exchange with science representatives and other doctoral candidates within the Volkswagen Group

Contact person for this posting: Fabian Wenzel