Senior Deep Learning Software Engineer

Job Description

The VPU team designs and delivers industry leading low power AI inference IP and software for edge use cases including Client and IoT platforms. The Compiler team is developing an NN Compiler for deployment of Deep Learning models on Intel Movidius hardware accelerator. The job of the compiler is to take an already trained neural network model, optimize the graph and format the model to a command list or compiler assembly code that can be run by the hardware accelerator.

Role’s responsibilities will include compiler design in addition to software development. The compiler features must be designed according to the hardware capabilities. The compiler is based on MLIR framework and utilizes its API and features (dialects, passes, etc.) to schedule and optimize the list of tasks which correspond to model inference.

Qualifications

• BS/MS in Computer Science Computer Engineering or a similar field
• At least 6 years of experience in programming
• C/C++ embedded programming
• Design of multithreading and multithreaded application
• Proven experience in low level optimization and vectorization

Preferred Qualifications:

• Proven experience in modern compiler architecture and code compilation process (e.g. llvm compiler, Java compiler)
• Theoretical and applied skills in Machine Learning
• Deep Learning frameworks (Tf, Caffe, PyTorch, etc) is a plus