PeopleofSuperintelligence

Researcher

Ronan Collobert

As of October 2026, Ronan Collobert is a research scientist at Apple Machine Learning Research, after research posts at Facebook AI Research, Idiap Research Institute, and NEC Laboratories of America. He earned a pure-mathematics master's degree from the University of Rennes in 2000 and received a machine-learning PhD from the University of Paris VI in 2004 after graduate studies at Idiap and the University of Montreal. He and Samy Bengio published the SVMTorch paper on large-scale SVM regression in 2001; he also coauthored Torch, a modular machine-learning software library, in 2002. With Jason Weston, he published a deep neural network paper on multitask learning for natural language processing in 2008; it won ICML's Test of Time Award in 2018. He worked on wav2letter in 2018, helped develop MLX in 2023, and coauthored a self-distillation paper for code generation in 2026.

Now at
Apple
Works
168
Citations
34,566
h-index
52

Publication numbers from OpenAlex, as of 15 October 2026.

Timeline

  1. 2000–2002

    A degree, SVMTorch, and Torch

    He earns a master's degree in pure mathematics from the University of Rennes in 2000. With Samy Bengio, he publishes SVMTorch, a paper on large-scale SVM regression, in 2001. In 2002, he and Bengio publish the Torch technical report with J. Mariéthoz; it describes a modular machine-learning software library.

  2. 2004

    A PhD in machine learning

    He pursues machine-learning graduate studies at the University of Montreal and Idiap and receives his PhD from the University of Paris VI in 2004. His thesis focuses on large-scale learning.

  3. 2008

    A multitask NLP paper

    With Jason Weston, he publishes a deep neural network paper on multitask learning for natural language processing in 2008.

  4. 2018

    A Test of Time Award

    In 2018, he and Jason Weston receive ICML's Test of Time Award for their 2008 paper.

  5. 2018

    Facebook speech recognition

    In 2018, he works as a Facebook research scientist on wav2letter, an end-to-end lightweight speech-recognition system.

  6. 2023

    Building the MLX framework

    In 2023, he and Awni Hannun, Jagrit Digani, and Angelos Katharopoulos initially develop MLX, an array framework for Apple silicon.

  7. 2026

    Self-distillation for code

    In 2026, he coauthors “Embarrassingly Simple Self-Distillation Improves Code Generation.” The paper's method raises Qwen3-30B-Instruct's pass@1 score on LiveCodeBench v6 from 42.4% to 55.3%.

  8. As of October 2026

    Apple machine learning research

    As of October 2026, he is a research scientist at Apple Machine Learning Research. His prior research posts include Facebook AI Research, Idiap Research Institute, and NEC Laboratories of America.

Sources and credits

Photo: We have no published photo of Ronan Collobert yet. Send us a free-licensed photo.

  1. Natural Language Processing From Scratch - Microsoft Research microsoft.com. Checked 15 October 2026.
  2. SVMTorch: Support Vector Machines for Large-Scale Regression Problems jmlr.org. Checked 15 October 2026.
  3. Ronan Collobert ronan.collobert.com. Checked 15 October 2026.
  4. ICML 2018 Awards icml.cc. Checked 15 October 2026.
  5. Facebook researchers win Test of Time award engineering.fb.com. Checked 15 October 2026.
  6. GitHub - ml-explore/mlx: MLX: An array framework for Apple silicon github.com. Checked 15 October 2026.
  7. Embarrassingly Simple Self-Distillation Improves Code Generation - Apple Machine Learning Research machinelearning.apple.com. Checked 15 October 2026.
  8. Publications - Apple Machine Learning Research machinelearning.apple.com. Checked 15 October 2026.