Your mission :
The postdoctoral scientist will work in a recently awarded collaborative project involving the International Committee of the Red Cross (ICRC), ETH and the Qatar Computing Research Institute (QCRI). In this project, we will work towards highly resolved building characterization and populations estimation from very high-resolution satellite images and social media data. Among other methodologies, the project involves extensive deployment of deep learning technology. Several African cities will be the target of the project.Main duties and responsibilities include :

  • Development of deep learning models for delineating buildings from very high resolution satellite images and extracting relevant parameter
  • Development of fusion models able to inject data from social media for better buildings estimation
  • Collaboration with the QCRI and ETH partners towards precise local population density estimation
  • Development of a demonstrator of the technology to be handled to ICRC at the end of the project

Your profile : 

  • PhD degree in computer science and/or environmental science
  • Proficiency in programming in Python / C++
  • Track of knowledge of deep neural networks technology; web and database programming
  • Knowledge of remote sensing and satellite data processing
  • Spoken / written proficiency in English and French
  • You are motivated, curious and willing to work in a highly dynamic team

We offer : 

  • An opportunity to develop a scientific career in an exciting area of science
  • A unique opportunity to learn advanced scientific and engineering techniques and approaches
  • Excellent educational conditions and competitive remuneration
  • A multi-cultural and stimulating scientific environment, in the new EPFL-Valais campus in Sion

Start date :
March 1st, 2021Term of employment :
Fixed-term (CDD)Duration :
1 year, renewable for a second year.Contact :
For additional information, please contact Prof. Devis Tuia by e-mail to


Leave a Reply

error: Alert: Content is protected !!
%d bloggers like this: