PhD Position: Physics-Informed Generative AI for Synthetic Energy Data

Radboud University is seeking a PhD candidate to join the NWO-funded SHARE project, focusing on the development of generative AI models for synthetic energy data. The research aims to overcome privacy and regulatory barriers in the energy sector by creating realistic, privacy-preserving datasets that adhere to physical power-flow equations. You will work with real-world operational data from Alliander to build tools for grid planning and decision-making. The role involves collaborating with an interdisciplinary team and contributing to an open-source synthetic data toolbox. You will also have the opportunity to publish at top machine learning venues and engage in limited teaching activities.

• Design and evaluate deep generative models including VAEs, GANs, and diffusion models for energy time series.
• Integrate physical constraints, such as Kirchhoff's laws and operational bounds, into generative architectures.
• Develop graph neural network models to represent complex network topologies.
• Create validated benchmark datasets and evaluation frameworks for statistical and physical fidelity.
• Collaborate with researchers on integrating differential privacy into the generative pipeline.
• Contribute to an open-source synthetic data toolbox for use by energy sector practitioners.
• Participate in teaching activities for computing science programmes up to 10% of the time.