Jobid=f40821ff0689 (0.0156)
Job Overview
Join the ARGOS project team to develop a novel end‑to‑end AI framework enabling soil moisture estimation from raw SAR data onboard satellites.
Job Description
Challenge
Spaceborne Earth Observation (EO) systems are critical to address Europe’s environmental and security challenges, from climate change monitoring and disaster response to border surveillance and situational awareness. Synthetic Aperture Radar (SAR) satellites are particularly valuable given their ability to capture data independent of weather or lighting conditions. However, the complexity of SAR signals, the computational demands of image formation, and the limited resources of spaceborne platforms hinder the deployment of advanced onboard applications. The ARGOS project aims to overcome these bottlenecks.
Change
The ARGOS project (Artificial Intelligence for Real‑time Guidance of Onboard SAR applications) will develop a novel end‑to‑end Artificial Intelligence (AI) framework enabling high‑level SAR applications directly from raw data onboard satellites. The ARGOS project will design a unified onboard processing chain capable of handling SAR raw data and delivering near real‑time application‑specific inferences in different fields. To make these processes viable in the constrained space environment, ARGOS will focus on optimizing deep learning models, applying multi‑tasking and Tiny AI approaches to reduce computational load and power consumption. The framework will be validated on representative space‑qualified hardware, ensuring that the solutions can operate effectively given realistic resources. In parallel, the AI‑driven approach will be benchmarked against state‑of‑the‑art onboard SAR processing methods. The project will also demonstrate its versatility through demonstration scenarios, including maritime situational awareness, critical areas identification, and environmental monitoring.
Impact
ARGOS will advance Europe’s autonomy in intelligent EO capabilities, enabling faster, more efficient, and resilient responses to environmental and geopolitical challenges. Ultimately, the project will bridge the gap between algorithmic innovation and operational deployment, serving both environmental and security needs while reinforcing Europe’s leadership in space‑based intelligence.
What you’ll do
TU Delft will lead the work package focused on the definition and consolidation of the system requirements, onboard applications and demonstration scenarios across a range of application areas. TU Delft will directly define the requirements and consolidate the use cases related to soil moisture estimation. In addition, TU Delft will contribute to the development of the E2E onboard AI framework for soil moisture estimation in applications where real‑time response is critical.
Responsibilities
- Generate synthetic and real‑world datasets used as input for E2E DL architectures. This will include SAR imagery and derived products.
- Label and structure datasets to support supervised training of E2E AI models.
- Establish a validation and testing strategy for the E2E AI models for soil moisture estimation based on community best practices.
- Validate estimated soil moisture against reference remote sensing datasets and in‑situ measurements.
- Be responsible for the curation, storage and maintenance of reference datasets for training and validation of the E2E onboard AI framework for soil moisture estimation.
- Contribute to the definition of the requirements, and consolidation of the use cases related to soil moisture estimation.
- Contribute to reporting and project management, including participation in project meetings and workshops.
- Publish your research results in peer‑reviewed publications.
- Present your research outputs in project meetings, workshops, and international conferences.
Working arrangement
You will join the Department of Geoscience and Remote Sensing at TU Delft (Faculty of Civil Engineering and Geosciences) and work closely with Prof. Susan Steele‑Dunne, her team at TU Delft (m-wave.tudelft.nl) and international collaborators across the ARGOS project team.
Qualifications and Experience
- A PhD in Earth Observation, Geodesy or a closely‑related field.
- Demonstrated experience with SAR image processing using SNAP, PolSARPro or similar.
- Experience in large‑scale validation of EO products against in‑situ data would be an asset.
- Demonstrated proficiency with Python or similar scientific programming environment.
- Ability to work with large datasets, develop reproducible workflows and apply modern data science tools.
- Strong organizational skills and a collaborative mindset: the ability to work effectively with international collaborators.
- Excellent communication skills and the ability to write scientific publications independently.
- Good command of written and spoken English.
Conditions of employment
- Duration of contract is 1 year.
- A job of 38–40 hours per week.
Salary and benefits
Salary and benefits are in accordance with the Collective Labour Agreement for Dutch Universities. The TU Delft offers a customisable compensation package, discounts on health insurance, and a monthly work costs contribution. Flexible work schedules can be arranged.
Diversity and inclusion
At TU Delft, we value diversity as one of our core values and actively engage to welcome all individuals, promoting inclusivity and equality across the university community.
#J-18808-Ljbffr
Deel deze vacature: