Freelancer profile translated to English.
Back to original languageAbout Romain
R&D engineer specialized in computer vision and deep learning, with experience covering the entire chain: from applied research to deployment in production on constrained targets.
I work on imaging problems where precision truly matters: 3D medical imaging (segmentation, anatomical landmark detection, registration), real-time rendering (super-sampling and neural denoising for path tracing), and 3D reconstruction/scanning.
I don't just train existing models: I modify the architecture to fit the problem. Adding specialized heads, replacing or adapting Transformer blocks, redesigning attention mechanisms, modifying filter chains, I adapt the network to the constraint (precision, latency, memory) rather than the other way around. I design, train, optimize, and deploy.
Technical skills:
Computer Vision: detection (YOLO, FPN), segmentation (nnU-Net), landmark detection, hybrid CNN + Transformer architectures (Swin, BiFormer, ViT), medical image registration (Elastix, VoxelMorph)
Deep Learning: PyTorch, architecture design and modification, training, unified inference pipelines
3D Imaging: medical volumes (CT/MRI), point clouds, depth sensors (Orbbec)
Deployment & Optimization: TensorRT, quantization, edge inference (Jetson), Modal, containerization
Full-stack to produce complete solutions: Python, React / Next.js, Node.js, APIs, cloud integration (Vercel, Supabase, Firebase)
I prefer to understand what needs to be solved before coding. Documented architecture choices, tested code, real-world deployment.
French
Native or bilingual
English
Fluent
Remote only
Primarily works remotely
Experience
- AMDR&D AI DeveloperFebruary 2026 - Today (5 months)Montpellier, FranceResponsible for research, modification, and optimization of existing Deep Learning models. Direct application to improving NSSD and developing real-time denoising solutions for video games.
- Zimmerbiomet,Python/C++ Developer (AI Division)June 2022 - February 2026 (3 years and 8 months)Montreal, QC, CanadaI specialized in semantic segmentation of anatomical bones (scapula, humerus, pelvis, etc.) with in-depth mastery of 3D and 2D UNet architectures. More recently, I dedicated the last year to innovative research projects, focusing on 3D registration, anatomical landmark prediction (Heatmap regression model), and 3D morphing. These works require expertise in reading and understanding scientific articles, as well as mastery of deep learning frameworks like PyTorch and VTK/ITK.
- AutologArtificial Vision DeveloperFebruary 2021 - May 2022 (1 year and 3 months)Montreal, QC, CanadaDevelopment of AI models and inference integration for defect recognition in wood materials.
Recommendations
Be the first to recommend Romain
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 Preparation in Medical ImagingETS Montreal2016Préparation au PhD en Imagerie Medicale
- Master 2Université Paul Sabatier (Toulouse III)2015Master 2