Stellenangebot / Job offer
Postdoc position in the area of Model-Driven Machine Learning
The Model-driven Machine Learning group at Helmholtz-Zentrum Hereon is seeking a curious and driven candidate to develop new machine learning tools for accelerating fluid simulations. The desired starting date is 1. Oct 2022 or shortly thereafter, and the position is initially limited to 3 years.
This project aims to develop machine learning algorithms that learn from numerical simulations to integrate fluid flow PDEs at a fraction of the cost of classical computational methods. A particular focus will be hybrid methods, where a neural network is used to improve or initialize a classical iterative method. Application domains include atmosphere and ocean simulations in cooperation with experts at Hereon and University of Hamburg, as well as simulations of planetary mantle convection developed by our project partners at the German Aerospace Center (DLR) in Berlin.
The ideal candidate will have expertise and hands-on experience in developing numerical methods for fluid dynamical PDEs. Strong mathematical and programming skills are essential, while machine learning expertise is desired but not a requirement. Experience with discretizations and integration schemes of atmosphere or ocean models is welcome but not necessary.
The Model-driven Machine Learning group combines ML- and physics-based approaches to Earth system modeling, within the Institute of Coastal Systems – Analysis and Modeling at Helmholtz-Zentrum Hereon in Geesthacht, near Hamburg, Germany. We combine flexible remote work with frequent in-person interactions. The successful candidate will focus on challenging and impactful research, with minimal administrative or teaching duties and supported by extensive on-site expertise in ML and Earth science, and top-notch computational resources.
Your tasks:
- develop hybrid methods that combine deep learning with classical numerical methods for fluid dynamical PDEs physical models describing ocean and atmospheric physics
- absorb and synthesize knowledge from machine learning, numerical modeling and Earth science
- write, publish and present your research in scientific journals and conferences
Your profile:
- PhD in mathematics, physics, machine learning, computer science or the geosciences
- can carry out research independently with curiosity, creativity, initiative and persistence
- works well in a small group with diverse skill sets and backgrounds
- learns quickly on the job and from research articles on ML and the geosciences
- solid publication record showing creative ideas and hard work. High-profile journal articles are welcome but not required.
- solid mathematical foundations including linear algebra, multivariate calculus, fluid dynamical PDEs
- comfortable programming in Python. FORTRAN, HPC or ML framework experience is welcome
- communicates well in written and spoken English
