Organisationseinheit

Institut für Theoretische Informatik (ITI)

Ihre Aufgaben

Do you want to pursue a PhD in Computer Science focused on the development of deep learning methods and apply these methods to a critical climate risk factor societies will face in the 21st century? If so, we encourage you to apply for our PhD position within the multidisciplinary, international research training group C4LaNd!

You will focus on advancing probabilistic deep learning models for spatiotemporal forecasting, with the primary goal of modelling multiple dimensions of wildfire risks, using data from both computer simulations and Earth observations (e.g., satellite data). A central component will be the development of improved algorithms, building on recent advances in AI such as versions of diffusion models. You will then apply these methodological developments to support climate change risk assessments. For example, your project will explore the combined impacts of climate change and changes in land use on biodiversity.

Ultimately, your wildfire model is intended to become a component of the dynamic global vegetation model LPJ-GUESS and of the AI world model WOW of the Earth system . As such, your work will also lead to an improved online representation of wildfires in future climate and ecosystem simulations.

Background: C4LaNd is a new research training group that brings together senior and doctoral researchers from the natural, social, economic and engineering sciences to address the challenges of sustainably using land for the benefit of both people and nature. You will receive a dual PhD degree from KIT and University of Melbourne, including a one-year research stay in Melbourne. Further information on C4LaNd, and the Ph.D. positions advertised can be found at C4LaNd.earth.

This PhD project will be based in the AI in Climate and Environmental Sciences research group at the Computer Science department of KIT in Karlsruhe, Germany. Your primary PhD supervisor will be Prof. Peer Nowack, and you will collaborate closely with Prof. Almut Arneth and Dr. Carolina Natel from KIT’s Atmospheric Environmental Research institute. Your Melbourne co-advisor will be Dr. Benjamin Henley.

Lines of research include:

• Advancing deep learning methods for probabilistic spatiotemporal modelling.
• Using these methods to assess multiple dimensions of wildfire risk under future climate and land use scenarios, in particular concerning impacts on biodiversity.
• Collaborating with fellow C4LaNd PhD researchers and with modellers in the LPJ-GUESS and WOW research communities.
• Reviewing literature and other data sources to support your work.
• Presenting your results at national and international conferences.

Eintrittstermin

November 2026

Ihre Qualifikation

• A Master’s degree in computer science, mathematics, physics, data science, or a related field.
• Strong programming skills, ideally in Python and C++.
• Fluency in English (written and spoken).
• Delight in engaging within an international, multidisciplinary training programme.

• Willingness to undertake a one-year research stay at the University of Melbourne.