The A. Mathis Group for Computational Neuroscience and AI (CNAI) is seeking two postdoctoral fellows to join our team working at the intersection of artificial intelligence and neuroscience. Our research is built around three intertwined questions: how to measure and understand behavior with AI, how brains and embodied agents learn to control the body, and how the brain builds a sense of its body through proprioception. We develop widely-used open-source tools (e.g., DeepLabCut), train biomechanically realistic embodied agents (e.g., MuscleMimic, Kinesis, Arnold), and build AI-based models of sensorimotor processing.

We are recruiting two postdocs:

1) Postdoc in Computer Vision & AI for Behavior Analysis

2) Postdoc in Embodied AI (Reinforcement Learning for Motor Control)

Main duties and responsibilities

For both positions:

• Conduct innovative, independent research aligned with the lab's mission and the candidate's interests
• Contribute to the lab's open-source software, datasets, and benchmarks
• Mentor PhD and Master's students and contribute to the lab's collaborative culture
• Engage with collaborators across EPFL and beyond (neuroscience labs, ecology partners, clinical collaborators)
• Present research at international conferences and workshops

Position 1: Computer Vision & AI for Behavior Analysis

• Contribute to the creation of behavioral datasets for groups of animals
• Develop methods for animal and environment reconstruction
• Develop methods for behavior understanding from multi-view video and audio recordings
• Advance multimodal large language models for fine-grained action understanding and reasoning
• Help shape the next generation of DeepLabCut and related open-source tools used by labs worldwide.

Position 2: Embodied AI

• Train muscle-actuated, biomechanically realistic agents to perform skilled motor tasks using reinforcement learning, imitation learning, and curriculum methods;
• Develop new methods for full-body musculoskeletal control
• Probe links between learned control policies and neural representations of movement in collaboration with experimental labs

Common requirements:

• PhD (completed or close to completion) in computer science, applied mathematics, computational neuroscience, robotics, physics or a closely related field;
• Strong publication record in relevant venues;
• Solid programming skills in Python and a modern deep-learning framework
• Excellent communication skills in spoken and written English;
• Self-motivated, collaborative, and committed to open science and reproducible research;
• Demonstrated ability to work both independently and as part of a team.

More details and ways to apply: https://careers.epfl.ch/job/Gen%C3%A8ve-Postdoctoral-Positions-in-Computational-Neuroscience-&-AI/1164982155/

Lab website: https://cnai.epfl.ch/