Master Thesis - Metric-Guided Diffusion Models for High-Fidelity In-Cabin Synthetic Data Generation - REF5866V

Are you looking for an exciting master’s thesis at the intersection of Generative AI, Computer Vision, and Automotive AI? Would you like to work with state-of-the-art diffusion models while contributing to privacy-safe, next-generation driver monitoring systems?

Then this thesis could be the perfect fit for you.

We are offering a 6-month master’s thesis in our AI laboratory in Berlin, with a preferred start date of February 2026. 

Your Thesis Tasks:

As part of this master’s thesis, you will:

• Implement and experiment with diffusion-based image generation models
• Develop an evaluation pipeline for a large-scale, high-fidelity synthetic in-cabin image dataset
• Investigate metric-driven feedback loops to improve image generation quality
• Analyze the impact of synthetic data on downstream computer vision tasks
• Summarize findings in a master’s thesis and potentially a research paper or patent submission

While not mandatory, this thesis is especially suitable for students considering a future PhD or research-oriented career.