About the position

Darwin Geospatial is looking for a candidate to join a three-year Industrial PhD programme developed in collaboration with Universidad Rey Juan Carlos .

The selected candidate will work on SoilSaver , an applied R&D; project focused on developing artificial intelligence systems to detect, characterise and predict soil erosion, land degradation and desertification using Earth observation data.

A central objective of the PhD will be the creation of a large-scale, public, multimodal and reproducible geospatial dataset designed for the training and evaluation of machine learning models.

The candidate may also contribute to other Darwin Geospatial R&D; activities related to habitat mapping, urban green infrastructure, ecological connectivity and environmental monitoring.

Key responsibilities

- Design and build the SoilSaver dataset by integrating Sentinel-2 imagery, PNOA orthophotos, digital elevation models, LiDAR data, climate variables and reference information.
- Develop machine learning and deep learning models for the detection, semantic segmentation and temporal prediction of erosion and land degradation processes.
- Build reproducible pipelines for data preprocessing, model training, validation and evaluation.
- Explore semantic segmentation architectures, geospatial foundation models, graph neural networks and physics-informed approaches.
- Analyse satellite image time series and environmental variables.
- Contribute to the publication of open datasets, code, scientific papers and conference presentations.
- Support the deployment of research outputs into Darwin Geospatial’s operational tools and products.
- Collaborate on additional R&D; projects involving automated habitat mapping, urban green infrastructure, ecological connectivity and environmental monitoring.

Essential requirements
- An official Master’s degree,



or equivalent qualification, granting access to doctoral studies.
- Academic background in engineering, computer science, geomatics, remote sensing, environmental sciences or a related discipline.
- Demonstrable experience in the design, training, validation and evaluation of machine learning models , preferably deep learning models.
- Solid knowledge of Earth observation and remote sensing data .
- Experience working with multispectral imagery, georeferenced raster data, spectral indices or satellite time series.
- Strong proficiency in Python and relevant data analysis, geospatial processing and machine learning libraries.
- Ability to develop and document reproducible analytical pipelines.
- Fluent spoken and written English. C1 level is recommended.
- Ability to write technical documentation and communicate results in an international scientific environment.
- Analytical thinking, autonomy and motivation to work on applied research.

What we offer
- A three-year employment contract with Darwin Geospatial.
- Development of a doctoral thesis in collaboration with Universidad Rey Juan Carlos.
- Gross annual salary between €22,000 and €31,000 , depending on experience and suitability for the position.
- Joint supervision from both the university and the company.
- Direct involvement in applied R&D; projects with potential integration into operational geospatial products.
- Access to cloud infrastructure, GPU resources and high-performance computing systems.
- Opportunities to contribute to scientific publications, international conferences, open datasets and technology development.
- Work at the intersection of artificial intelligence, geospatial data, ecology and environmental management.

How to apply Interested candidates should send an updated CV to:

[email protected]

Recommended email subject

Industrial PhD Application – "Full name"

📌 Industrial PhD in Remote Sensing and ML for Soil Degradation (Madrid)
🏢 Darwin Geospatial
📍 Madrid