Description
CEA is hiring a six-month internship in Saclay, France, focused on cross-architecture knowledge distillation for embedded systems. The intern will study and implement methods that transfer knowledge from Vision Transformer or Swin models to compact ResNet or MobileNet networks, including designing alignment strategies, projectors, adapters, and loss functions, and evaluating the results on target hardware using the Aidge framework in C++ and Python. The role requires a Master’s year 2 or equivalent, C++ and Python skills, and deep learning knowledge; embedded systems development knowledge is preferred.
