Jean Mercat
Jean Mercat John

Senior Research Scientist

About Me

I’m a senior machine learning research scientist at the Toyota Research Institute, where I specialize in transformers, large language models, vision language models, and large behavior models. My passion for ML has led me to explore a variety of fields, including self-driving cars, robotics, and language processing. I care about a careful scientific process including thorough evaluation-driven experimentations, I aim for a large downstream impact of my research, and I want to always learn from awesome coworkers.

Download CV
Interests
  • Artificial Intelligence
  • Natural Language Processing
  • Multi-modal Language Models
Education
  • PhD Machine Learning

    Paris Saclay University, L2S and Renault

  • MEng in Scientific Computing

    ENSEIRB-MatMéca, Bordeaux, France

📚 My Research

I’m a senior research scientist at Toyota Research Institute. I pre-train, uptrain, fine-tune, experiment, and do research with Large Language Models, Vision Language Models, and Large Behavior Models.

I attempt to understand and improve large models, their evaluation process, and their training data. I apply large models to robotic manipulation, and agents to push the boundary of open-ended embodied intelligence.

Featured Publications
Recent Publications
(2025). A Careful Examination of Large Behavior Models for Multitask Dexterous Manipulation. arXiv preprint.
(2025). OpenThoughts: Data Recipes for Reasoning Models. arXiv preprint.
(2025). Should VLMs be Pre-trained with Image Data?. In ICLR 2025.
(2024). DataComp-LM: In Search of the Next Generation of Training Sets for Language Models.
(2024). Linearizing Large Language Models. In COLM.
Recent & Upcoming Talks