Martin Chatton
AI Software Engineer at Deeplink.ai
Lausanne, Vaud, Switzerland
Summary
🤩
Rockstar🎓
Top SchoolMartin Chatton is an AI Software Engineer based in Lausanne with 12 years of experience building production-ready ML systems and scientific software. Trained at EPFL (BSc CS, MSc Data Science), he has moved between industry roles from data science at a commodities firm to ML engineering in sports and now AI engineering at Deeplink.ai, bringing a pragmatic bridge between research and product. He contributes to open-source projects like pyopencl and StarDist, improving numerical robustness and test coverage for scientific computing and object-detection pipelines. Comfortable with low-level performance details and high-level ML model behavior, he pays attention to edge cases and reproducibility—evident from tests for inplace operations, NMS, and class-aware predictions. Mart in combines hands-on coding, test automation, and domain-aware modelling to deliver reliable, well-tested AI components.
12 years of coding experience
3 years of employment as a software developer
High School Diploma, Applied Mathematics and Physics, High School Diploma, Applied Mathematics and Physics at Gymnase de Beaulieu, Lausanne, SwitzerlandMaster's degree, Data Science, Master's degree, Data Science at EPFL (École polytechnique fédérale de Lausanne)
French, English, German
Github Skills (14)
object-detection10
opencl10
computer-vision10
machine-learning10
pytest10
python10
numpy10
testing10
parallel-computing9
arrayobject9
gpu8
data-analysis8
performance-monitor7
performance-analysis7
Programming languages (9)
Github contributions (5)
stardist/stardist
Jul 2018 - Jun 2022
StarDist - Object Detection with Star-convex Shapes
Role in this project:
Data Scientist & ML Engineer
Data Scientist & ML EngineerContributions:4 releases, 17 reviews, 525 commits in 4 years
Contributions summary:Martin contributed to the development and testing of the StarDist object detection library. They implemented a non-maximum suppression test and improved the handling of edge cases with constant images. Furthermore, they added a test case for the accurate detection of object instances using the model on a 2D example image and integrated surface calculations into the rays object. Additionally, the user integrated new test cases and made it possible to use the model with a custom set of classes for the multi class prediction head.
pytorchconvexobject-detectionstardistbioimage-analysis
inducer/pyopencl
Jan 2017 - Feb 2019
OpenCL integration for Python, plus shiny features
Role in this project:Backend Developer & Test Automation Engineer
Backend Developer & Test Automation EngineerContributions:10 commits, 1 PR, 7 comments in 2 years
Contributions summary:Martin primarily focused on enhancing the `pyopencl` library by implementing and testing inplace division functionality for the `Array` class. They added tests for inplace division with both scalars and arrays, ensuring compatibility with NumPy's behavior. Furthermore, the user addressed potential type casting issues and code formatting, demonstrating a commitment to code quality and accuracy within the OpenCL integration.
amdarraypythonreductionopencl
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.