Stage - Développement et intégration de modèles VLA pour la robotique autonome - Industrie - Toulouse
Core
Design and integrate Visual Language Action (VLA) models to enable autonomous robots to perceive, reason, decide, and execute missions in real-world environments.
Role type
Intern, Robotics AI Engineer (VLA)
Builds
Autonomous robot systems with perception, reasoning, and action capabilities
Domain
Robotics, Artificial Intelligence, Embedded Systems
Deliverable
production ML models
Required skills
Python, C++, ROS 2, VLA model training, real-time system constraints, simulation-to-real transfer, model optimization for embedded hardware
Preferred skills
Experience with NPU/CPU acceleration, asynchronous execution patterns, semantic scene representation
Technologies
ROS 2, Python, C++, NPU, CPU
Responsibilities
Define semantic interfaces between perception and VLA reasoning; design and train VLA policies for mission execution; integrate VLA into ROS 2 architecture; implement real-time asynchronous execution; optimize VLA models for embedded deployment; validate policies via simulation and real-world testing
