54141: (Geo-) Computer Scientist, Data Scientist, Statistician or similar (f/m/x) - Developing methods to ensure the quality of information in AI systems for Earth observation

Area of research:

Scientific / postdoctoral posts

Job description:

The department “Data Science for Earth Observation” develops modern signal processing and AI methods for current and future Earth observation missions. We are involved in a large number of third-party funded projects and are part of a large international network.

Your tasks will consist of scientific research on the topic of artificial intelligence for Earth observation as well as the initiation, implementation and management of scientific and innovative projects.

Your area of responsibility extends to research and development of procedures and methods to ensure the quality of information in AI systems related to Earth observation. The tasks described below require specific knowledge and outstanding expertise. In order to meet the requirements of the position, extensive professional or research experience in the respective fields is required.

Tasks include:

• literature research with the aim of critically evaluating new experimental and theoretical methods, making them usable for our own work or evaluating the problems, difficulties and disadvantages that arise, deriving new approaches and solutions
• independent research and expansion of scientific knowledge at the highest international level in the field of uncertainty assessment, interpretability and traceability of neural networks
• review and further development of existing methods for robustness analysis and uncertainty modeling of machine learning models. New and further development of approaches for (visual analysis) of model components or for self-explanatory models
• research and development of procedures for the verification and stabilization of machine learning models, e.g. through prior knowledge or through the combination of multiple data sources
• perform experiments to validate the developed methods. Evaluation and interpretation of the experimental results to compare different methods
• implementation of the procedures in prototypical algorithms
• detailed preparation and presentation of the new approaches to scientific users
• presentation and discussion of the research results at conferences, meetings, etc. at national and international levels
• publication of the research results in suitable publications
• participation in topic-related scientific committees, working groups and interest groups. Representation and presentation of the scientific results internally and externally
• provide guidance for doctoral students as part of the research project

This research center is part of the Helmholtz Association of German Research Centers. With more than 42,000 employees and an annual budget of over € 5 billion, the Helmholtz Association is Germany's largest scientific organisation.