Students and applicants
Opportunities
Research topics for M1, M2, and PhD students.
Prospective topics
PhD
Efficient architectures for vision and language models; small models for visual document analysis; continual learning in the era of multimodal foundation models.
Prospective applicants are welcome to make contact to discuss future possibilities.
M1 projects (TER: 8 weeks / 1 day per week at the lab)
Offers for Paris-Saclay students only.
(M1 - TER) · filled · [VLMs][Evaluation]: building an evaluation dataset for visual document analysis of historical scientific documents.
M2 internships (6 months)
Contact me for details (sending M1 transcript and M1 internship report if any).(M2 - internship) [Frugality][Evaluation] Automatic leaderboard generation to analyse the energy efficiency of ML models.
(M2 - internship) [ML4Physics] Neural Operators for solving partial differential equations.
Previous openings (2025/2026)
Small language models: efficient adaptation and continual evaluation
CIFRE in partnership with SCIAM.
Thesis director: Sophie Rosset.
Summer 2026 research internships at LISN
Topic 1: Multimodal continual learning.
Topic 2: Benchmarking parameter-efficient fine-tuning methods for spoken-language understanding.
Location: LISN laboratory, Université Paris-Saclay, Orsay, France.
Supervisors: Eva Feillet and Sahar Ghannay.
Summer 2026 research internships at LISN
Topic 1: Visual continual learning.
Topic 2: VLMs for cultural heritage.
Location: LISN laboratory, Université Paris-Saclay, Orsay, France.
Supervisors: Eva Feillet.