CS27 - Bac+5 Stage Deep Reinforcement Learning pour la Prise de Décision en Conduite Autonome (H/F)
Core
Develop a Deep Reinforcement Learning (DRL) system for autonomous highway driving decision-making, specifically optimizing lane changes in a simulated environment.
Role type
Research intern (Deep Reinforcement Learning)
Builds
Simulated autonomous driving decision systems
Domain
Autonomous driving / Robotics / AI
Deliverable
production ML models
Required skills
Python, Deep Learning, Reinforcement Learning, PyTorch, TensorFlow, Stable-Baselines3, Reward Engineering, Simulation environments
Preferred skills
Autonomous driving domain knowledge, Control theory
Technologies
highway-env, DQN, PPO, SAC
Responsibilities
Analyze state-of-the-art scientific literature on RL for autonomous driving; Configure and model traffic scenarios in highway-env simulation; Design, implement, and train DRL algorithms; Define and tune reward functions for safety, comfort, and efficiency; Evaluate model performance against complex scenarios and benchmark against classical ADAS systems.
