The Role:
The Accelerator Architecture team focuses on VEC/MATRIX architecture and potentially CPU/Scalar architecture. Specifically, the team will define the VPU and MPU architecture, along with all micro-architecture choices, using performance, power, and area analysis. This will involve providing the SoC team with design guidelines tailored for the specific application.
Key responsibilities:
As an Accelerator AI Architecture Engineer, you will be instrumental in the research, development, and verification of cutting-edge VPU (Vector Processing Unit) and MPU (Matrix Processing Unit) architectures, with a strong emphasis on software-hardware co-design principles. Your core responsibilities will involve working on advanced device architectures, defining novel designs for matrix/tensor units, architecting the next generation of high-performance AI accelerators, and ensuring optimal interaction between hardware and software. This includes overseeing design definition, implementation, and rigorous verification, as well as developing comprehensive design examples and sophisticated simulation models, all supported by detailed technical documentation.
In this role, your responsibilities will include, but are not limited to:
- Research and develop novel micro-architectures and designs for high-performance matrix/tensor processing units, with a focus on enabling efficient software execution.
- Innovate and architect the next generation of high-performance AI accelerators, considering the interplay between hardware capabilities and software requirements.
- Drive software-hardware co-design efforts to optimize overall system performance and efficiency for AI workloads.
- Support the development of optimized architectures capable of significantly improving PPA (Performance-Power-Area) while effectively integrating customer requirements and considering software implications.
- Evaluate the architectural implementation of both VPU and MPU designs, ensuring they meet performance targets, power constraints, and facilitate efficient software utilization.
- Support the seamless integration of the VPU and MPU with the scalar CPU, effectively addressing and resolving data transfer bottlenecks, while also considering the software interface and programming models.
Qualifications:
- 5+ years of solid experience in CPU, MPU (Matrix Processing Unit), or SoC design, with a strong emphasis on accelerator architectures and software-hardware co-design.
- Proven experience in the research and development of matrix/tensor unit designs for AI acceleration, with a clear understanding of software implications.
- Deep understanding of various scalar and vectorial processor architectures, with specific expertise in architectures relevant to AI and machine learning, and how software interacts with them.
- Strong working knowledge of the architectural challenges in multi-core and many-core accelerator design, including software parallelization and programming.
- Familiarity with RISC-V architecture and its application in accelerator design, considering both hardware and software aspects.
- Extensive working knowledge of functional and performance simulators used for architecture exploration and verification, including the ability to model software workloads.
- Solid understanding of processor micro-architecture, with a keen ability to analyze and optimize performance, power, and area trade-offs in the context of AI accelerators and their software ecosystem.
- Experience in architecting high-performance AI accelerators for various applications, with a strong emphasis on software programmability and performance.
- Familiarity with common AI/ML software frameworks (e.g., TensorFlow, PyTorch, ONNX) and how they map onto hardware architectures.
Soft Skills:
- Teamwork
- Excellent communication
- Engagement
- Innovation
- Customer oriented
We are looking for outstanding people willing to join our mission to change this industry and help to build a better world.
If you feel identified with Openchip, please contact us. We can offer a competitive compensation package in a flexible work schema that will help you to keep a balance between your personal and professional life.
At Openchip & Software Technologies S.L., we believe a diverse and inclusive team is the key to groundbreaking ideas. We foster a work environment where everyone feels valued, respected, and empowered to reach their full potential – regardless of race, gender, ethnicity, sexual orientation, or gender identity
