Automated individual-level data collection based on photos for using deep learning on a wild bird

Background
Individual-level data is crucial when working in evolutionary biology and is mostly performed by capturing and marking animals with tags. Recent computational advances, specifically deep learning, are now opening new possibilities to collect data on animals in a less invasive way.

We recently implemented procedures for automating the collection of training data (pictures), generating training datasets for AI models with the aim of allowing to individually identify birds (Ferreira et al 2020 MEE) and extract other features from the birds based on images. We worked with Sociable weaver, a monomorphic endemic bird from southern Africa.

The aim of this internship is work with us to set up a system in the field that takes pictures from multiple angles, which will generate additional training data to help us improve our current AI models and develop new ones. During this internship the student will have the opportunity to further develop fieldwork skills with other RFID systems.

Location of the internship: CEFE-CNRS University of Montpellier and / or CIBIO University of Porto

Duration internship: 3 months, starting in march 2026. We preferably look for student in gap year

Timeline: Early march: training in Switzerland with André Ferreira; Mid march-end of may: fieldwork in South Africa (Benfontein reserve near Kimberley) including more training in place at the beginning.

Work done in the field: (i) set and maintain the system to take photos and move it between colonies every day (one day off a week), (ii) download and checking the data (iii) help running two other tasks already implemented in the field : an automatic scale and RFID feeders that are related with the setup used to collect the pictures.

Competence needed: Some experience working in the field in remote places. Interest or experience working with electronics for data collection (e.g. RFID, raspberry pi). Experience working with R or Python.

Requirement. Having a driving license and being able to speak in english with a South African team

What is covered: Travel to the field. Field accommodation and insurance paid; Stipend: 600 euros per month, Food: not covered

Field condition. You will work in a team that at minimum included 2 field managers.

More information: https://sociableweaverproject.com/about/

NB : If the position is not field in march 2026, it will re-open in septembre 2026 for 6 months and with training back in the lab to learn about deep learning data and models