I am passionate about exploring the applications of machine learning and artificial intelligence for a sustainable live within our planetary boundaries.
I am an AI expert at Birds on Mars, where I am responsible for AI applications for environmental sustainability. I am member of the Board of Directors at Climate Change AI, an international nonprofit catalyzing impactful work at the intersection of climate change and machine learning.
I have been an external lecturer on AI and data science among others at TU Berlin, Leuphana University Lüneburg, the Climate Change AI summer school and CODE University. Previously, I was a research associate at the TU Berlin, where I led the Smart Energy Systems research group of the DAI Lab, working on AI applications in the smart grid and sustainable development of AI systems. I like to give talks about AI and sustainability in different contexts, e.g. at Bits and Bäume, a key note at the DESSAI Inria-DFKI European Summer School on AI 2022 or the Bitkom Big-Data.AI summit 2022.
If you are working on machine learning and smart meter data, check out our recent open access book on load forecasting, this Python tutorial, and our list of load datasets.
PhD in Computer Science, 2023 (planned)
TU Berlin
M.Sc. in Information Systems, 2014
Humboldt University of Berlin
B.Sc. in Information Systems, 2011
HWR Berlin
Is the first textbook on load forecasting for the distribution network. Brings together both statistical and machine learning topics. Includes colorful illustrations and practical examples from many sectors and developing countries. Is open access, which means that you have free and unlimited access.
The transition to a fully renewable energy grid requires better forecasting of demand at the low-voltage level to increase efficiency and ensure reliable control. However, high fluctuations and increasing electrification cause huge forecast variability, not reflected in traditional point estimates. Probabilistic load forecasts take uncertainties into account and thus allow more informed decision-making for the planning and operation of low-carbon energy systems. We propose an approach for flexible conditional density forecasting of short-term load based on Bernstein polynomial normalizing flows, where a neural network controls the parameters of the flow. In an empirical study with 3639 smart meter customers, our density predictions for 24h-ahead load forecasting compare favorably against Gaussian and Gaussian mixture densities. Furthermore, they outperform a non-parametric approach based on the pinball loss, especially in low-data scenarios..
This paper presents a literature review on the topic of Low Voltage (LV) load forecasting. It gives an overview of the approaches, core applications, datasets, trends, and challenges. Suggestions how to facilitate the continued improvement and advancement are given and a set of recommendations toward best practices are provided.