William Peoc'h
Machine Learning Research · AI for Science
University of Oxford ML research intern working on physical inverse problems, multimodal learning, and biological data.
Available for a six-month final internship beginning in January 2027.
Research Focus
Machine Learning for Scientific Data
Machine learning researcher and engineer working on AI for Science, physical inverse problems, multimodal learning, and biological data. Currently an ML research intern with the Hallou Group at the University of Oxford, developing models that infer per-cell stress tensors from microscopy images. Available for a six-month final internship beginning in January 2027.
Research Experience
Oxford, Inria, and applied machine learning work for scientific and biomedical data.
Machine Learning Research Intern
- Developing deep learning methods for a physical inverse problem: inferring per-cell stress tensors directly from microscopy images of cellular tissues.
- Designing and systematically evaluating local-global architectures that combine pretrained vision encoders with cross-attention to fuse cell morphology and tissue-level context.
- Building a reproducible PyTorch experimentation pipeline spanning synthetic data generation, distributed multi-GPU training on Isambard-AI, controlled ablations, and model diagnostics.
Machine Learning Research Intern
- Fine-tuned transformer models for biomedical text classification, named-entity recognition, and information extraction using PubMed-derived datasets.
- Built NLP experimentation pipelines covering dataset construction, preprocessing, model training, evaluation, and error analysis.
- Conducted reproducible experiments on HPC infrastructure, including model comparisons, hyperparameter studies, and controlled ablations.
Machine Learning Intern
- Developed CNN and U-Net models for retinal image classification and segmentation across heterogeneous imaging datasets.
- Integrated the inference pipeline into a C# application used by ophthalmologists for retinal image analysis.
Selected ML Projects
Projects selected for research depth, machine learning internals, and original modeling work.
Multimodal Medical Assistant
Built at the Mistral AI x Alan Hackathon: a multimodal medical assistant combining image analysis, conversational AI, and voice interaction. Awarded First Place and Best Pitch.
First Place and Best Pitch - Mistral AI x Alan Hackathon, 2024
Education & Selected Award
Academic background, followed by selected recognition.
Education
MSc in Bioinformatics and Modeling
Research-focused curriculum in machine learning, applied mathematics, statistics, computational biology, and numerical modeling.
Exchange Semester – Mathematics & Machine Learning
Completed exchange semester in mathematics and machine learning.
Two-year degree in Computer Science
Two-year computer science program at the University of Reims.
Selected Award
Mistral AI x Alan Hackathon – 1st Place & Best Pitch
Mistral AI · Alan
Built an AI medical assistant with image diagnostics and a voice/chat interface; ranked 1st among 150+ participants and received the Best Pitch Award.
Contact Me
Available for a six-month final internship beginning in January 2027.
Contact Information
Languages
French
Native
English
C1 – TOEIC 900/990
Spanish
B1
Chinese
A2
Research Focus
Final Internship Search
I am looking for machine learning research and research engineering roles in AI for Science, physical inverse problems, multimodal learning, or biological data.