We launched recruiters.rip, let AI apply for you. First month free with code onemonthfree

Research Internship – Physical AI

Paris-based non-profit open-science AI research lab developing and releasing multimodal AI models, demos, code, and research.

6 current maintainers5 active leads
Paris, France

Funding history

About Kyutai

Kyutai is an independent, non-profit AI research lab focused on open science. Its current work includes real-time speech and voice systems, vision-speech models, language models, and world models; it releases research artifacts and maintains interactive demos.

View jobs by Kyutai

Skills

Candidate Availability

Required and preferred rules are kept separate and reflect the wording in the original posting.

About the Role

You will join a team of experienced researchers and engineers on a 6-month internship in Paris, with a flexible start in 2027. Together with your supervisor you will define a project in areas such as world models, vision-language-action models, motion retargeting and reinforcement learning. You will review prior work, form testable hypotheses, build and train models, run experiments at scale, validate in simulation and deploy on real robots, and contribute to papers and open-source releases.

Requirements

  • Master's student (M2) or final-year student at an engineering school or university in machine learning, robotics, applied mathematics or computer science
  • Strong competence in machine learning, robotics or robot learning
  • Solid in math and algorithms and able to turn theory into working code
  • Fluent in Python and a deep learning framework (PyTorch or JAX)
  • Fluent with AI agents for coding and research
  • Curious and rigorous, able to work in a team

Responsibilities

  • Define your research subject with your supervisor and have it approved by your school or university
  • Review relevant literature and reproduce key results
  • Form sharp, testable hypotheses by critically analyzing existing approaches
  • Design architectures and training methods, and build or curate datasets
  • Run experiments at scale and evaluate against baselines
  • Validate work in simulation and deploy it on real robots, closing the sim-to-real gap
  • Contribute to research papers and release code and models on GitHub or Hugging Face

Benefits

  • Hands-on access to real robotic platforms
  • Significant compute for experiments
  • Possibility to publish, attend scientific conferences and release open-source code
  • Option to continue with a PhD at Kyutai
  • Competitive internship allowance and benefits