This position is for a Deep Learning Software Engineer in AIPG’s Movidius core NN team, supporting upcoming VPU IP and SOC products. Your responsibilities will include:
- Implementation of embedded run time scheduling for hardware.
- Design and code of distributed real-time embedded inference kernels, designed and optimized for a specific hardware architecture, implementing both basic and cutting edge network features for multiple types of numerical precision.
- Profiling and optimization of layer and full network operation.
- Work in both pre and post silicon environments
- Unit test, numerical accuracy and performance verification of work
Come join our IAA and industry award winning team!
Qualifications
Requirements:
- Must have either a BS or MS in Computer Science, Computer Engineering or similar field
- 5 years hands-on coding experience in modern C++ programming language
- Significant DSP, embedded RTOS or bare metal coding and debugging experience
- Numerical coding experience in mixed precision such as float16, bfloat16, int8 and other models
- Familiarity with Deep Learning frameworks (TF, Caffe, PyTorch, OpenCV, etc.)
- Development experience in a Linux environment
The following is highly desired:
- Previous Deep learning layer implementation experience
- Modern compiler architecture and back-end coding experience
- Host side GPU shader workload memory management and/or scheduler experience
- Spoken and written English: upper-intermediate level or advanced