Advanced Engineer - High-Efficiency AI Computing

About the Company

We collaborate with a globally reputed leader in information and communication technology (ICT) infrastructure and smart devices, delivering comprehensive full-stack, all-scenario solutions for products and services spanning carriers, enterprises, governments, and individual consumers worldwide.

Role Overview and Key Responsibilities

As an Advanced Engineer specialized in High-Efficiency AI Computing, you will lead efforts to innovate AI hardware technologies. Your role is crucial in integrating algorithmic breakthroughs with silicon design to advance low-precision and sparse computing capabilities. The outcome of your work will directly shape the microarchitecture of our flagship AI accelerators, substantially improving performance and energy efficiency for large-scale AI applications.

• Leading research on cutting-edge quantization techniques for complex low-precision data formats to expand our low-precision computing ecosystem and maintain a leading edge in numerical computing.
• Designing and implementing high-efficiency, scalable computational kernels (such as GEMM and FlashAttention) optimized specifically for our proprietary AI accelerator microarchitecture.
• Undertaking hardware-software co-design tasks by detecting systemic performance constraints and spearheading the microarchitectural enhancement of next-generation AI chips to boost throughput and energy conservation.
• Collaborating intensively with integrated circuit design teams by delivering detailed technical specifications, comprehensive benchmark documentation, and valuable patent applications.

Qualifications

• Ph.D. degree in Computer Science, Electronic Engineering, Automation, or related advanced technical disciplines.
• Solid foundational knowledge in computer architecture with particular expertise in GPU and NPU microarchitecture implementation.
• Extensive expertise and proven experience in algorithmic research focusing on low-precision quantization and sparse computing.

Preferred Attributes

• Demonstrated success in optimizing inference for Large Language Models or multi-modal AI models along with practical experience developing high-performance kernels.
• Hands-on system-level architectural design experience for high-throughput AI inference accelerators.
• A notable academic publication record in prestigious architecture and AI conferences such as ISCA, MICRO, HPCA, ASPLOS, NeurIPS, and CVPR.

Additional Information

Applications are welcome from qualified candidates who meet the criteria outlined. Only shortlisted applicants will be contacted. By submitting your application and personal data, you consent to the collection, use, and disclosure of your information by PERSOL Singapore Pte Ltd and affiliates as detailed in their Privacy Policy. EA License No.: 01C4394.