About Weiyuan
ML & Computer Vision Engineer | Python, PyTorch | PhD
Je conçois et déploie des pipelines de vision par ordinateur et de machine learning en Python, depuis l’analyse d’images jusqu’au déploiement de modèles deep learning en production.
J’aide les équipes à transformer des images ou des datasets complexes en systèmes automatisés fiables.
Ce que je peux faire pour vous :
• Détection, segmentation et classification d’images (OpenCV, PyTorch)
• Conception de pipelines d’analyse d’images robustes et reproductibles
• Entraînement et déploiement de modèles deep learning
• Nettoyage et structuration de datasets images (qualité, doublons, biais)
• Automatisation de workflows data et création d’outils internes (APIs, CLI)
Expérience
• 8 ans d’expérience en R&D image et data (CNRS, Aix-Marseille Université)
• Auteur de imgclean, outil open-source de nettoyage de datasets images (PyPI)
• 3 publications scientifiques premier auteur (80+ citations)
Stack
Python • PyTorch • OpenCV • NumPy • C++ • SQL • Docker • CI/CD
Domaines
MedTech • Biotech • Vision industrielle • Startups ML
Disponible pour missions de computer vision, analyse d’images ou prototypage ML.
French
Native or bilingual
English
Fluent
Can work on-site
Strasbourg (up to 50km)
Experience
- N/AIndependent ML / CV EngineerTECHSeptember 2025 - Today (9 months)Strasbourg, FranceOpen-source tools & applied ML projects• • Published imgclean on PyPI — a CLI toolkit for automated image dataset auditing (duplicates, blur, corruption, train/val leakage detection). 50 automated tests, CI/CD, CLIP-based semantic analysis.• • Built CineDepthPro, an on-device deep learning portrait engine for Android using monocular depth estimation and custom GPU shaders, achieving real-time 4K rendering.• • Developed simu, a C++20 orbital simulation library with RK4/Verlet integrators, conjunction detection, CMake packaging, and Python bindings.
- Le laboratoire « Matière et Systèmes ComplexesR&D Engineer – Computer Vision & Imaging SystemsRESEARCHOctober 2021 - August 2025 (3 years and 10 months)Paris, France• • Built and deployed deep learning pipelines (U-Net) processing 100 k+ microscopy images, reducing manual analysis cost by 50% and enabling reproducible quality control.• • Designed an Arduino/JavaScript real-time control system synchronising sensor acquisition with confocal imaging, replacing a manual workflow.• • Created automated data validation pipelines improving measurement robustness across heterogeneous imaging datasets for a 10-person research team.
- The Marseille Developmental Biology InstituteR&D Engineer – Scientific Imaging & Quantitative AnalysisRESEARCHOctober 2017 - July 2021 (3 years and 9 months)Marseille, FrancePhD in Developmental Biophysics• • Developed a force-inference library adopted by international teams; reimplemented in Python as an open-source package (Bayesian solvers, graph-based topology, OpenCV).• • Published 3 first-author papers (80+ citations); built a C++/Qt laser-control interface for image-guided micro surgery.
Recommendations
Be the first to recommend Weiyuan
Help this freelancer shine by sharing your experience working together.
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- PhD in Developmental BiophysicsAix-Marseille University2021PhD in Developmental Biophysics
- Master's Degree –University of Strasbourg2017Master's Degree –