R&D Engineer in Artificial Intelligence for avifauna monitoring (F/M/X)

The objective of the position is to contribute to the development of artificial intelligence algorithms dedicated to the automatic detection of birds from video systems and camera traps. The candidate will work closely with the AI and avifauna teams of FEM’s Biodiversity & Interactions department, and will play a central role in the implementation, training and evaluation of deep neural networks. The work will be based on annotated datasets as well as on various data processing and deep learning model training tools developed in the framework of the DRACCAR-MMERMAID project.

In this context, the candidate will be required to:

• To carry out a literature review on neural network architectures adapted to object detection tasks in video streams, with a particular interest in the tracking of birds in offshore environments.
• Contribute to the definition of learning and evaluation strategies, as well as to the preparation and structuring of training, validation and test datasets.
• Participate in the implementation of the selected neural network architectures, as well as in the development of the associated learning algorithms and evaluation protocols.
• Analyze and interpret the performance of models to propose ways of improvement (architectures, data, training strategies).
• Contribute to the