Ronan Collobert
Research Scientist at Apple
Menlo Park, California, United States
Summary
🤩
Rockstar🎓
Top SchoolRonan Collobert is a research scientist with 15 years of experience applying and advancing machine learning across industry labs, currently at Apple and previously at Meta and NEC Labs. His work spans core ML systems and CUDA-accelerated deep learning, with significant open-source contributions to Torch (torch7, cutorch, cunn, nn) and the C++ Flashlight library where he implemented robust multi-threaded BlobDataset support. He combines deep research background (IDIAP/Université de Montréal PhD-era roles) with hands-on systems engineering, often bridging algorithmic ideas and high-performance implementations. Ronan is comfortable working close to hardware—optimizing CUDA kernels and tensor backends—while also improving usability and testing across libraries. Colleagues can expect a focus on correctness, performance and maintainability, evidenced by his threading, serialization and API compatibility fixes in widely used ML toolkits. Based in Menlo Park, he brings a scientist’s rigor to production-grade ML infrastructure.
15 years of coding experience
21 years of employment as a software developer
Lycee Felix Le DantecPierre and Marie Curie University
Github Skills (34)
pytorch10
c-language10
operation10
tensorrt10
gpu-programming10
machine-learning10
c1110
ml10
c1710
lua10
deep-learning10
tensorflow10
arrayfire10
neural-network10
cuda10
Programming languages (8)
Github contributions (5)
torch/torch7
Jan 2012 - Jan 2017
http://torch.ch
Role in this project:
Back-end Developer & Library Maintainer
Back-end Developer & Library MaintainerContributions:145 commits, 14 PRs, 11 pushes in 5 years
Contributions summary:Ronan primarily contributed to the development and maintenance of the core library functions within the torch7 repository. Their commits demonstrate a focus on adding new mathematical operations, such as logical operations on tensors, and optimizing existing functions like `normall` by leveraging atomic operations for thread safety. Additionally, the user addressed bugs and improved the robustness of the system. This included fixing threading and serialization issues.
libtorchtorchc-plus-plustensorautograd
torch/nn
Jan 2012 - May 2015
Role in this project:Backend Developer
Backend DeveloperContributions:52 commits, 4 PRs, 2 pushes in 3 years 3 months
Contributions summary:Ronan primarily focused on modifying and enhancing the core functionality of the `nn` package within the `torch/nn` repository. Their work included implementing new features like `TemporalMaxPooling`, refactoring and improving existing modules such as `SpatialConvolutionMM`, and addressing various warnings and code cleanup tasks. Additionally, the user addressed several bug fixes and standardized the codebase through integration of the new luaT API.
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.