About Sarah
With over four years of experience in physics and artificial intelligence, I currently work as a research engineer at the Icube/GAIA Platform laboratory at the University of Strasbourg on the innovative Helpmewalk project. My expertise focuses on the application of machine vision, deeplearning, 3D reconstruction and complex physical systems, acquired in particular during my PhD at the IJL in partnership with APREX Solutions.
I have strong experience in fast imaging, visual tracking, anomaly detection and 3D spatial reconstruction from sensor and image data, applied to challenging environments such as nuclear fusion plasmas and industrial systems.
I collaborate with multidisciplinary teams to design and deploy advanced AI-based perception and modelling solutions, bridging experimental physics, computer vision and data-driven modeling, while promoting innovation and scientific excellence.
My goal is to contribute to projects that combine 3D perception, physics-aware AI and real-world applications to address major societal and technological challenges.
French
Native or bilingual
English
Fluent
German
Basic
Remote only
Primarily works remotely
Experience
- ICube,Plateforme GAIA, Université de Strasbourg,Research engineerBIOTECHMarch 2025 - March 2026 (1 year)Strasbourg, France• - Developed a deep-learning method PINNs based to estimate position and orientation of magnetic sensors from measurement data; implemented training/inferenceworkflows in Python/PyTorch with CUDA GPU acceleration, tracked experiments/metrics, versioned artifacts, and logged/shareable models with MLflow (model packaging + registry).• - Medical project: built a 3D shape reconstruction pipeline for custom orthosis design (pre-processing, segmentation, surface/mesh reconstruction, quality control) using OpenCV, scikit-image, Open3D; integrated outputs into clinical prototyping tools (CloudCom pare/ParaView).• - Applied teacher–student distillation to compress YOLO models for embedded/edge constraints.• - Performed simulation/experiment coupling and validation using Radia magnetic-field simulations; ensured reproducibility and maintainability with Git (clean code, version ing, documented pipelines).
- InstitutJeanLamour (IJL),PhD studentOctober 2020 - September 2024 (3 years and 11 months)Nancy, France• - and supervised deep-learning models for detection/segmentation of turbulent structures in ultra-fast sequences; built end-to-end data pipelines (pre-processing, labeling strategy, evaluation metrics, model monitoring).• - Developed unsupervised / anomaly-detection approaches to characterize electric arc defects and bubble dynamics, including feature extraction and clustering/representation learning; leveraged Anomalib and custom Python tooling.• - Implemented computer vision and diagnostic calibration workflows (e.g., Calcam, Ax Vision) and maintained reproducible analysis codebases across experiments.
- InstituePlasmaPhysicsResearch InternOctober 2021 - November 2021 (1 month)IPP, Greifswald, MV, GermanyExperimental/data-analysis work on COMPASS tokamak diagnostics; processed experimental datasets and contributed to interpreta tion/validation of analysis results.
Recommendations
Be the first to recommend Sarah
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 student in Plasma Physics and Machine VisionUniversity of Lorraine, Institut Jean Lamour2024PhD student in Plasma Physics and Machine Vision
- Polytech Nancy2020