Benoit Steiner
Benoit Steiner co-authored the papers that describe the TensorFlow and PyTorch deep learning systems. He completed engineering studies in France and then spent about ten years in electronic design automation. In 2016 he was one of the Google Brain authors of the TensorFlow paper, and in 2017 he co-authored a paper on device placement with reinforcement learning. He co-authored the PyTorch paper at NeurIPS in 2019, and he took part in an ML for Systems panel at MLSys in 2022. He focuses on systems research at Facebook.
Timeline
-
By 2016
Engineering studies, then EDA work
He completes engineering studies in France at École Centrale Paris and computer science at École des Mines de Paris. About ten years in electronic design automation follow, and by 2016 he works at Google Brain.
-
2016
The TensorFlow system paper
TensorFlow: A System for Large-Scale Machine Learning appears in the OSDI 2016 proceedings, and he is among its co-authors.
-
2017
Device placement with reinforcement learning
In 2017 the proceedings of the 34th International Conference on Machine Learning publish Device Placement Optimization with Reinforcement Learning, with him as a co-author.
-
2019
The PyTorch library paper
At NeurIPS 2019, PyTorch: An Imperative Style, High-Performance Deep Learning Library appears with him as a co-author.
-
2021
ML for Systems workshop
He appears in the opening remarks of the ML For Systems workshop at NeurIPS in 2021.
-
2022
A panel on ML for systems
A panel at MLSys in 2022 discusses the practical adoption of ML for systems in industry, and he is one of its panelists.
-
2023
Memory and compiler papers
In 2023 he co-authors MODeL: Memory Optimizations for Deep Learning, and a paper on learning compiler pass orders, both in the proceedings of the 40th International Conference on Machine Learning.
-
By 2025
Systems research at Facebook
In 2025 he focuses on systems research at Facebook.
Sources and credits
Photo: We have no published photo of Benoit Steiner yet. Send us a free-licensed photo.
- About - Benoit Steiner
- TensorFlow: A System for Large-Scale Machine Learning | USENIX
- TensorFlow: A system for large-scale machine learning
- Device Placement Optimization with Reinforcement Learning
- PyTorch: An Imperative Style, High-Performance Deep Learning Library
- NeurIPS Opening Remarks
- MLSys Panel Discussion
- MODeL: Memory Optimizations for Deep Learning
- Learning Compiler Pass Orders using Coreset and Normalized Value Prediction