Hackathon on
Machine Learning for the Earth System
27-29 August 2025
Bonn, Germany
We are pleased to announce a new Hackathon on Machine Learning for the Earth System for PhD students and early-career researchers. This event is associated with our third Workshop on Machine Learning for the Earth System Modelling right before the hackathon.
What is a Hackathon & Why Machine Learning for Climate Science?
A hackathon is an event where people come together to solve real-world problems through coding, collaboration, and innovation. It’s a fast-paced, hands-on experience where teams brainstorm, develop, and present projects within a short timeframe.
Machine Learning (ML) is revolutionising many fields, including weather and climate science. From improving numerical weather predictions to developing climate emulators and foundation models, ML helps us better understand and predict the Earth system. This hackathon provides a unique opportunity to apply cutting-edge ML techniques to critical challenges in weather and climate research.
What is this about?
Join us for an exciting Hackathon on Machine Learning for the Earth System! This event brings together PhD students and early-career researchers from Earth System science and Computer Science to collaborate, learn, and tackle challenges at the intersection of AI and climate research.
📅 Date: August 27–29
📍 Location: Computer Science Department, University of Bonn – Friedrich-Hirzebruch-Allee 8, 53115 Bonn, Germany
💰 Fee: 50 € (includes access to the venue, extended coffee breaks and lunch for the 2.5 days)
The application procedure will open in mid March. Visit the registration page for details.
Stay updated—subscribe to our mailing list at mlesm-workshop@listen.uni-bonn.de to receive announcements!
What is ahead for you?
Why Join?
-
Inspiring Talks – Kickstart with expert-led technical sessions.
-
Hands-on Coding – Work on real-world challenges in weather and climate modeling.
-
Mentorship & Networking – Learn from and connect with researchers and developers.
-
Team Collaboration – Whether you’re a beginner or an expert, find teammates and build together.
-
Exciting Prizes – Compete for awards in different tracks!
What to Expect?
Building on the momentum of workshops and research initiatives in deep learning for the Earth system, this hackathon will provide a collaborative event that brings together developers, researchers, and enthusiasts from Earth System science and Computer Science. Participants will tackle critical challenges and learn about large-scale deep learning in weather and climate.
This hackathon features two tracks:
-
Beginners Track – Learn the basics, collaborate with mentors, and build impactful projects. No advanced experience required—just basic knowledge of ML (e.g., RF, FFN). Take part in guided lectures and hands-on exercises.
-
Advanced Track – Dive deep into complex challenges, explore cutting-edge ML techniques, and develop innovative solutions.
Potential project topics include:
-
ML-based numerical weather predictions (NWP)
-
Climate emulators
-
Foundation models for Earth System Science
Example project: Use the open-source Weather Generator code—or bring your own project! Mentorship will be available to guide you.
Participants will have the opportunity to attend expert-led lectures covering a range of topics in Earth System Science (ESS), Machine Learning, and general computational practices.
In the ESS domain, lectures will introduce fundamental concepts of weather and climate forecasting, working with ESS data, and evaluating weather and climate models. Additionally, participants will explore current research topics in weather and climate science.
For ML, sessions will cover an introduction to Graph Neural Networks, fine-tuning and utilizing pre-trained models, and coding attention mechanisms for weather AI models. Participants will also learn about training and evaluating weather AI models on multiple GPUs.
Beyond these specialised areas, there will also be a lecture on best practices in open-source software development.
Is this for me? ✅
✔ I am a PhD student, early-career researcher, or ML enthusiast.
✔ I have an interest in AI and climate science.
✔ I want to gain hands-on experience in ML for Earth sciences.
✔ I enjoy collaborating with peers and experts in the field.
✔ I am eager to learn and contribute to impactful projects.
This hackathon is for you! 🎉 Whether you’re an experienced developer, researcher, or just someone passionate about learning more about AI this hackathon welcomes participants of all backgrounds and skill levels.
Teams can be formed during the hackathon. Don’t worry if you don’t know anyone—we’ll help you find a team.
Organising Committee
-
Daniel Minge (TRA1 Modelling, University of Bonn)
Questions?
📩 Contact us at: mlesm-workshop@cesoc.net
We gratefully acknowledge the support by
CESOC, ECMWF, the TRA Modelling at the University of Bonn, the Forschungszentrum Jülich and the University of Cologne.
Supported by
