Description

I am a research-trained physicist turned Machine Learning Engineer, with strong foundations in mathematics, programming, and image science, and over five years of hands-on experience delivering computer vision and ML systems in startup environments.

My background allows me to bridge research and production: I am comfortable designing and training advanced models, but also turning them into robust, scalable solutions integrated into real products. I have worked end-to-end, from data definition and model experimentation to cloud-based training, performance tracking, and deployment through APIs or containerized services.

I enjoy working on applied AI problems where scientific rigor meets engineering constraints, and where machine learning creates measurable impact in real-world systems.

Langues

Chinois
Bilingue ou natif
Anglais
Capacité professionnelle complète
Français
Capacité professionnelle limitée

Préférences en matière de lieu de travail

Accepte de travailler sur site

Palaiseau (jusqu’à 30 km)

DXOMARK
Machine Learning Engineer (ComputerVision)
janvier 2024 - juin 2025 (1 an et 5 mois)
Paris, France
A consulting firmspecializing in image quality assessment.
• • Design guidelines and prepare training data for image quality annotations.
• • Design machine learning pipelines from model training to production. Keywords: data analysis, model design (RESNET, VIT, SWIN Transformers, etc.), model deployment, image signal processing, color science.
Machine learning Computer Vision Transformers FastAPI Data analysis
OCUS
Data Scientist (ComputerVision)
septembre 2022 - janvier 2024 (1 an et 4 mois)
Paris, France
A middle-sizestartup that harnesses thepower of tech toproducesmart visual content and help brands and creators thrive.
• • Trained several high-accuracy classification models, e.g., watermarks, texts, and logos in one image (data preparation, guidelines definition, 3rd-party and manual annotation, and model fine-tuning).
• • Pilots owner: The product owner of experimental projects following the client's requirements. Keywords: model tuning & evaluation, algorithm design, API integration, OpenCV, Pytorch, Amazon EC2 & S3.
Computer Vision Transformers Python Machine learning AWS EC2
Flyinstinct
Senior Solution Scientist
juin 2020 - septembre 2022 (2 ans et 3 mois)
Paris, France
Astartup thatprovidessolutions based on computer vision technology to industrialproducts for airport runway FOD (foreign object debris) detection.
• • Introduced an instant segmentation model (i.e., Faster-RCNN), which reduces more than 70% system false-positive errors.
• • Delivered a 3D-vision based module (i.e., Orthophoto), which yields better detection performance.
• • Invented an approach to match images with different luminosity (more efficient than the state-of-the-art, e.g.,
histogram-matching or Gamma-correction), which improves 50% of object detection rate.
• • Designed a localization algorithm based on perspective transform, leading to an increase of the detection rate from < 10% to > 90%. Keywords: proof of concept, algorithm design, computer vision, object detection & segmentation, machine & deep learning, data analysis, computer networking.
Computer Vision Data analysis Python Machine learning image processing