CV

Education

• Ph.D. student, IMT Atlantique, 2024-2027
  • Research on diffusion models for text generation
  • Supervisor: Vincent Gripon, Bastien Pasdeloup, Axel Marmoret
• Engineering degree, IMT Atlantique, 2021-2024
  • Mathematical and Computational Engineering specialization
  • Machine Learning, Deep Learning, Computer Vision, Signal Processing, Optimization
• M.Sc. Engineering Sciences and Applications, University of Rennes 1, 2023-2024
  • Signal and image processing major
• Academic exchange, University of Technology Sydney, 2023

Work experience

• November 2024: Visiting Scientist
  • Sony Europe, Stuttgart, Germany
  • Research on discrete diffusion models for text generation
• April 2024 - October 2024: Research Intern
  • Criteo AI Lab, Paris, France
  • Studied diffusion models, trained large diffusion models for image background generation
  • Supervisor: Jaspreet Sandhu, Thibaut Issenhuth, Ludovic Dos Santos
• Summer 2023: Research Intern
  • BRAIn team, IMT Atlantique, Brest, France
  • Studied parameter-efficient fine-tuning methods for SAM and attention-based data augmentation techniques
  • Supervisor: Bastien Pasdeloup
• February 2022: Developer Intern
  • French Polynesia’s government IT Department, Papeete, French Polynesia
  • Development of a Web application for network management, automation of database migrations

Skills

• Computer Science
  • Python, SQL, Java, R, Julia, C++, Matlab, Jupyter
• Machine Learning
  • Numpy, Pandas, sklearn, statistical and probabilistic models
• Deep Learning
  • PyTorch, Tensorflow, Keras, CNNs and Vision Transformers training
• Hugging Face 🤗
  • tranformers, diffusers, tokenizers, datasets, accelerate
• Distributed operations
  • Hadoop, Spark, multi-gpu/multi-node training (PyTorch distributed, Accelerate, NCCL), SLURM
• Soft Skills
  • Agile methods, communication, applied research
• Languages
  • English (C1, IELTS 7.5), Chinese (B2.1, HSK3 equivalent, intensive summer study at Beijing Language and Culture University), basic Spanish

Publications

FICUS: Few-shot Image Classification with Unsupervised Segmentation

Lys, et al. (2024). "FICUS: Few-shot Image Classification with Unsupervised Segmentation." In 2024 32nd European Signal Processing Conference (EUSIPCO)(pp. 1791-1795). IEEE.

Map

See a cool map of places I visited for conferences, internships and studies