Yann Henon
Machine Learning Engineer at Axis Communications
Antibes, Provence-Alpes-Côte d'Azur, France
Summary
🤩
Rockstar🎓
Top SchoolYann Henon is a Machine Learning Engineer with 11 years of experience specializing in computer vision and deep learning, currently based in Antibes, France and working at Axis Communications. He has a strong track record of building and productionizing object detection models, contributing core implementations of RetinaNet and Faster R-CNN in both Keras and PyTorch. Yann’s open-source work includes notable improvements to the widely used keras-retinanet project and a PyTorch reimplementation, along with algorithmic contributions like Harris and Shi-Tomasi corner detectors to the JuliaImages library. His background spans research and applied roles at startups and R&D teams, blending rigorous test automation and documentation with hands-on model engineering. Colleagues value his attention to training pipelines, loss refinement, and practical preprocessing tricks (such as aspect-ratio preserving resize) that boost real-world performance. He pairs academic grounding from Monash University with a pragmatic, open-source-first approach to shipping robust vision systems.
11 years of coding experience
8 years of employment as a software developer
Monash University
Stackoverflow
4,221reputation
231kreached
38answers
2questions
classification
top-1%
tensorflow
top-1%
augmented-reality
top-5%
machine-learning
top-1%
image-processing
top-5%
keras
top-1%
Github Skills (32)
algorithm10
algorithms10
pytorch10
python10
image-processing10
classification10
machine-learning10
data-structure10
keras10
tensorflow10
retinanet10
neural-network10
object-detection10
data-structures10
faster-rcnn10
Programming languages (4)
Github contributions (5)
yhenon/pytorch-retinanet
Apr 2018 - Dec 2019
Pytorch implementation of RetinaNet object detection.
Role in this project:
ML Engineer
ML EngineerContributions:34 commits, 14 PRs, 40 pushes in 1 year 8 months
Contributions summary:Yann implemented the core components of a RetinaNet object detection model using PyTorch. Their contributions include defining the Pyramid Features Network (FPN), regression and classification models, and the anchor generation mechanism. They also worked on the training loop, including loss calculation, optimization, and evaluation, integrating the COCO dataset for object detection training and validation.
object-detectionpytorchretinanetpython
small-yellow-duck/keras-frcnn
Jan 2017 - Feb 2017
Role in this project:ML Engineer
ML EngineerContributions:59 commits, 10 pushes in 1 month
Contributions summary:Yann primarily contributed to the implementation of the Faster R-CNN model for object detection using Keras. Their contributions include defining the model architecture, loss functions, and methods for generating training data, specifically focusing on the RPN (Region Proposal Network) and classifier components. They also made improvements to the data augmentation pipeline and model training scripts.
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.