Internship title: Artificial Intelligence: Exploration and application of Foundation Models and Machine Learning Workflows for Earth Observation data

Machine Learning (ML) is a powerful tool to extract information out of large volumes of satellite data. ML Workflows are a novel and quickly developing approach towards EO-based user applications, addressing a variety of societal information needs. New trends in ML workflows emerge in a very fast pace. In particular the generation/availability of analysis-ready data/ data cubes and use of Foundation Models significantly increase the efficiency of ML-powered applications.

Earth Observation data generated by EUMETSAT is particularly suitable for ML applications, as they provide observation about the full earth system. Applying ML techniques to the EUMETSAT data would enable the development of a variety of applications relevant for EUMETSAT Member States and the Copernicus User communities. Potential applications include weather forecasting and detection of weather extremes, operational oceanography, climate change, early detection of environmental phenomena such as marine algae blooms, air quality issues and ocean biochemical composition.

The internship would in particular address two different aspects, namely:

• The provision of analysis-ready data/data cubes from selected data produced by EUMETSAT;
• The identification and application of suitable foundation models for Earth Observation data

These two strands of work will enable the setup of ML workflows within the Copernicus WEkEO Data Access system for selected use case(s) which is the final objective of this internship.

The intern will be placed in the Copernicus Programme Office.

Duties

The main tasks of the internship will be to:

• Identify machine learning trends in the field of Earth Observation and provide an overview of uses cases and workflows (e.g. using geospatial foundation models) of relevance to EUMETSAT;
• Familiarise with the EUMETSAT operated missions and WEkEO processing infrastructure;
• Assess the usability and applicability of the identified trends and workflows for EUMETSAT and define one or more use cases in the area of climate, weather, marine and/or atmosphere;
• Modify, adapt and implement selected Machine Learning workflows for one or more of the identified use cases on the WEkEO infrastructure;

• Develop an easy-to-navigate interface to demonstrate the use case(s), e.g. generating relevant Jupyter Notebooks;
• Document the use case(s), used AI/ML/FM techniques and related workflows for further dissemination in EUMETSAT;
• Formulate recommendations and lessons learned based on the work done, including on the processing of earth observation data in the cloud, and the parallel processing of large datasets in Jupyter Notebooks.
• In addition, the intern will support relevant day-to-day activities within the team.

Skills and Experience

• Have the ability to work effectively in English;
• Be computer literate;
• Be intrinsically motivated by Innovation, Earth Observation data applications and AI;
• Be able to work independently and collaboratively;
• Have the ability to take the initiative in researching ideas;
• Have the ability to collect, collate, conceptualize and present information clearly.

Additional specific requirements:

• Knowledge about Earth Observation data and its use;
• Programming experience, preferably in Python;
• Configuration and usage of virtual cloud computing environments;
• Knowledge in other general purpose programming languages e.g. R and experience in writing code and contributing to open source projects are advantageous.