- https://www.ovt.com/products/oax8000/
- https://www.ovt.com/products/oax4600/
- https://www.ovt.com/products/oa7100/
- Neural Network Processor (NPU) architecture, programming Model, simulation, verification, profiling and optimization
- Specification, review and verification of other IP blocks
- FPGA images and IO boards for pre-silicon verification and design exploration
- Boot ROM, bare metal BSPs, sample applications
- Support Inference Engine, Toolchain and ASIC teams on NPU
- Constantly improve productivity through automation in all areas
Skills:
- Strong ANSI C bare-metal design, implementation, optimization and debugging using GCC-like toolchains
- Strong processor architecture knowledge: ARM/RISC-V ISA, symmetric multiprocessing, cache architecture, SIMD, Vector Processing, custom instructions
- Strong software, hardware and FPGA debugging capability
- Good communication, collaboration, task estimation & reporting, CI/CD via tools such as Confluence, Git, JIRA etc
- Familiarity with Python for prototyping, demo, verification and/or automation
- OS: GNU/Linux, proprietary RTOS, Bare metal
The following will be advantageous:
- Detailed and comprehensive understanding of TinyML (e.g. TensorFlow Lite for MCUs), from training, quantization through to operator execution, optimizing for power, performance, and size.
- Ph.D/MSEE/BSEE in Electrical Engineering, Computer Science or Computer Engineering
- At least 5 years of commercial embedded system architecture and development experience.
- Experienced in fixed point computing implementation strategies, such as quantization, scaling, saturation and overflow handling.
- Design, implement, and maintain high-quality C/C++ functional & bit-Accurate Model for NPU operation simulation