Embedded Camera System Trainee (Computer Vision)

TL;DR

Embedded Camera System Trainee (Computer Vision): Developing and optimizing image processing and CV algorithms for automotive cockpit products with an accent on deep learning inference and system performance. Focus on implementing AVM algorithms, fisheye distortion correction, and model quantization for embedded platforms.

What you will do

• Accelerate deep learning inference on embedded platforms using NPU and GPU.
• Develop and deploy AVM (Around View Monitor) and innovative CV algorithms for automotive cockpit products.
• Implement image processing modules, including image stitching, fusion, and fisheye distortion correction.
• Optimize models using quantization, pruning, and operator adaptation.
• Debug and tune overall system performance to ensure brightness consistency and processing efficiency.
• Track the latest research in image processing and AVM technologies.

Requirements

• Strong proficiency in C++ and Python.
• Solid understanding of Computer Vision and Deep Learning principles.
• Experience with frameworks such as PyTorch or TensorFlow.
• Knowledge of embedded systems, including HAL, DSP, NPU, and GPU.
• Familiarity with model deployment tools like ONNX Runtime, Qualcomm QNN, or SNPE.
• Must be eligible for 2027 campus recruitment (China).