To support the deployment of our Computer Vision solutions for In-Car Monitoring but not only, you will work under the guidance of our industry experts to implement and optimize a preselected Deep Learning-based algorithm and deploy it on an embedded platform. You will demonstrate its visual performance in camera-based application. Some example algorithms include Body Pose prediction, Hand Gesture detection, etc.
Your job will be to understand the neural network architecture, implement its execution code and select the appropriate optimization techniques such as efficient use of vector computation, memory caching or any specific neural computation acceleration SDK available for the platform. Your input will be the Pytorch pre-trained CNN model and your target platform will be an Qualcomm, Intel based or similar embedded platform with specific hardware compute units (CPU, DSP, GPU).
Skills required:
- Experience with embedded software development and specific toolchains
- C coding skills
Nice to have:
- Basic understanding of Deep Learning concepts and experience with PyTorch or similar