Work environment, missions and activities
Recruited at INRAE, you will be with the Joint Research Unit “Territories, Environment, Remote Sensing and Spatial Information” (UMR TETIS-INRAE/CIRAD/AgroParisTech/CNRS) within the “Maison de la Télédétection” (literally translated “the house of remote sensing”!), in Montpellier, France, Earth.The objective of TETIS is to master the processes for mobilizing spatial information for agro-environmental research and for the sustainable management of natural resources and territories. TETIS employs a hundred staff members. It is composed of 4 research teams, a “training” collective and an engineering team. Each research team develops a key axis for the implementation of an integrated approach to the spatial information chain. A specific of TETIS is to combine research, training and public policy support activities. TETIS is also strongly involved in the Theia land data center for sharing and promoting of research products.You will be part of the ATTOS team (“Acquisition, Remote Sensing, Processing and Space Observations”) and the engineering team. Part of ATTOS’ research focuses on the analysis and processing of satellite images, for the monitoring and sustainable management of artificial and natural environments and agro-ecosystems in a context of global change. This work makes extensive use of machine learning methods, particularly deep learning, enabling to take advantage of the growing amount of Earth Observation (EO) data with an unprecedented scale-up.You will work on projects in connection with CNES (French space agency), IGN (French geographical institute), and other public french institutions (like universities and other research facilities). You will contribute to the development of technical solutions and thematic products, intended for a wide range of users.You will mobilize deep learning for the analysis and processing of EO images, in land use mapping (e.g. mapping France’s buildings and structures at metric scale using VHRS images) and also in many other exciting things like image restoration (we lead an ongoing project for CNES on cloud removal in optical images) or image enhancement (e.g. super resolution). You will implement innovative analysis methods, and deploy them on computing infrastructures dedicated to AI.You have a passion for software development, machine learning, and earth sciences. You have a practical experience of scientific computation problems on large datasets, using python in a Linux environment.
You will be more specifically in charge of
– Employ your skills in deep learning and programming to develop applications to train and apply models on large volumes of satellite images (Sentinel, Spot).– Contribute to the development of open-source tools that promote research work in deep learning applied to remote sensing.– Deploy and use the codes on national high performance computing platforms, for instance CINES and IDRIS, and the computing resources of INRAE and CNES. The Jean-Zay GPU cluster, hosted at IDRIS, has more than one thousand GPU NVIDIA V100, which enables to go large scale!
Training and skills required
– Recommended training: Master’s degree in computer science or equivalent.– Desired knowledge : Programming, machine learning, signal/image processing– Experience required: tensorflow/pytorch practice, comfortable with python and linux, English read and written. Appreciated experience: gitlab or github, docker, contribution to open-source projects.– Required skills: creativity, curiosity, initiative, critical thinking, collaborative logic.
Your quality of life at INRAE
By joining INRAE, you will benefit (depending on the type of contract and its duration) :– up to 45 days leave per year– support for parenting: CESU childcare, leisure benefits– skills development programs: training, career counseling– social support: counseling and listening, assistance and social loans– vacation and leisure services: vacation vouchers, discounted accommodation– sports and cultural activities– catering services
Please send your letter and resume to Rémi Cresson (remi (dot) cresson (at) inrae (dot) fr)Come join us!