Principal Architect - Large Model and Training System Performance Optimization

The Computing Data Application Acceleration Lab aims to create a leading global data analytics platform organized into three specialized teams using innovative programming technologies. This team focuses on full-stack innovations, including software-hardware co-design and optimizing data efficiency at both the storage and runtime layers. This team also develops next-generation GPU architecture for gaming, cloud rendering, VR/AR, and Metaverse applications.

One of the goals of this lab are to enhance algorithm performance and training efficiency across industries, fostering long-term competitiveness.

• Lead the architecture design of Ascend training products, driving the continuous evolution of architectural competitiveness.

• Analyze mainstream scenario requirements and industry technology trends for Ascend, introducing innovative technologies to ensure sustained leadership in architectural competitiveness.

• Identify requirements for MindX, AI frameworks, acceleration libraries, and chip hardware, building a robust software-hardware architecture for Ascend training to achieve ongoing commercial success.

• Collaborate with other departments/teams from Huawei’s global research centers to align on strategic goals

• Spearhead project planning and define the technology/product development roadmap to guide long-term innovation

The base salary for this position ranges from $121,000 to $230,000 depending on education, experience and demonstrated expertise.

• Master’s or PhD in Computer Science, Math/Statistics, with a focus on AI & Deep Learning.

• 5+ years of experience in architecting large-scale AI training systems or similar complex software-hardware integrated solutions.

• Excellent documentation skills for writing internal reports and/or publishing research papers. Effective communication skills for presentations to internal and external audiences. A proactive attitude with a strong ability to tackle challenges and adapt to evolving requirements and dynamic work environment.

• Working knowledge of AI accelerators or full-stack AI acceleration systems and Deep Reinforcement Learning.

• Hands-on experience with veRL or Ray for large-scale model training.

• Familiarity with processor architectures and relevant work experience, with hands-on expertise in designing and developing complex system software architectures, and experience in performance optimization on GPU/NPU or similar hardware platforms.

• Solid understanding of deep learning fundamentals, proficiency with the PyTorch framework, and practical experience in performance optimization using upper-layer distributed frameworks such as Megatron or DeepSpeed.

• Strong programming skills with proficiency in C/C++ and Python.

• Experience using performance analysis tools such as Nsight Systems, Nsight Compute, and DLProf.