CS27 Bac+5 - Stage Ingénieur Data/IA - Machine Learning pour la validation ADAS (H/F)
Core
Design a multimodal machine learning model to automatically detect and classify driving scenes encountered during test drives using embedded camera and CAN bus signals.
Role type
Junior machine learning engineer (ADAS validation)
Builds
Production ML models for driver assistance systems
Domain
Automotive / ADAS validation
Deliverable
production ML models
Required skills
Python, PyTorch, scikit-learn, XGBoost, OpenCV, feature extraction, temporal signal processing, video analysis, model training and optimization, real-time inference
Preferred skills
Time series analysis, embedded systems constraints
Technologies
Python, PyTorch, scikit-learn, XGBoost, OpenCV, CANape, CANoe, MF4/MDF, Bus CAN, ONNX, Matlab/Simulink, Git
Responsibilities
Instrument test vehicles with cameras and record CAN bus data; synchronize video and CAN signals; define scene taxonomy and perform semi-automated labeling; extract features from CAN and video; compare fusion strategies; train and optimize ML models; evaluate model metrics and robustness; port model to real-time inference on series vehicles.
Seniority
Junior, internship (6 months)