About the role#
We are looking for a Master's student to join our EI² team for a 16-week research internship. You will work directly with our scientists and data scientists to assess whether Earth observation foundation models can improve our current tree species classification pipeline in Quebec. This is a part-time role, approximately 20 hours per week, starting in January 2028.
What you'll need#
- Current enrollment in a Master's program in remote sensing, geomatics, forestry, machine learning, or a related field.
- Solid Python skills, specifically NumPy, pandas, and scikit-learn.
- Knowledge of machine learning, including classification and model evaluation.
- Experience with satellite data, ideally Sentinel-2 or Google Earth Engine.
- Basic GIS skills using tools like rasterio, geopandas, or QGIS.
- Strong rigor in evaluating results.
- Ability to work in English.
- Basic PyTorch knowledge, a background in forestry or vegetation, and French language skills are considered assets.
What you'll do#
- Get familiar with our tree species classification pipeline and reference data, including field plots and forest stand polygons.
- Build a rigorous benchmark comparing Tessera and AlphaEarth embeddings with our current results.
- Train and evaluate simple classifiers, such as random forests or small neural networks, on these embeddings.
- Measure performance against the amount of labeled data used, ranging from 5% to 100%.
- Test whether mixed forest stand polygons improve overall results.
- Report per-species results, with particular attention to rare species.
- Deliver clean, documented code integrated into our repository, a short final report, and a presentation to the team.
Location & details#
- Location: Montréal, Quebec, Canada.
- Work Modality: Hybrid.
- Term: Winter 2028.
- This is a paid internship.


