Jobid=2b0fd4ddcd8c (0.0997)
The energy transition is one of the most urgent challenges of our time. The electricity grid is under increasing pressure due to electrification, renewable energy growth and rising demand, and AI is key to solving it. As a Postdoctoral Researcher in AI-based Load Forecasting at Radboud University, you will be at the heart of this challenge. You will develop cutting-edge AI models that predict electricity consumption at both grid and individual building level, directly contributing to reducing network congestion in the Netherlands. Your work will have real-world impact: the models you develop will be tested and validated together with industry partners such as Alliander and Stedin and made openly available to the broader AI and energy community.
You will conduct research on advanced AI methods for short-term load forecasting as part of the FlexLab AI Innovation Lab, a collaboration between Radboud University, Alliander, Stedin, the Netherlands Organisation for Applied Scientific Research (TNO), Eindhoven University of Technology (TU/e), HAN University of Applied Sciences and several SME partners. FlexLab develops, tests and validates AI technologies for flexible energy management with the goal of tackling network congestion in medium- and low-voltage grids. Radboud University leads the load forecasting work package (WP4), and you will be the key researcher driving this work forward.
Your research will focus on developing and validating AI models for load forecasting in scenarios where current models fall short, such as extreme weather events, grid incidents and high variability in renewable energy. You will explore techniques including graph neural networks, Bayesian neural networks, conformal prediction intervals and generative AI for synthetic data generation. You will also develop frameworks for uncertainty quantification in forecasting and integrate your models into the open-source OpenSTEF platform and the Linux Foundation Energy ecosystem.
You will work closely with SME partners in short-cycle innovation trajectories of 6 to 18 months, translating scientific advances into practical prototypes to be tested in real or simulated environments. You will contribute to scientific publications, open-source releases and knowledge-sharing events with the broader energy and AI community.
Deel deze vacature: