Co-op Researcher - AI Computing System

About the role

The Co-op Researcher will focus on advancing performance, efficiency, and usability of AI systems on the Ascend platform. Work involves low‑precision training, multimodal optimization, reinforcement learning, and training resource optimization to address system bottlenecks and deliver next‑generation AI capabilities.

Key responsibilities

• Design and develop optimization solutions for AI training and inference, including FP8 optimization and RL‑driven training agents.
• Combine AI algorithm requirements with system‑level architectural optimization in computing, I/O, scheduling, and precision control.
• Build stable, efficient AI training clusters using dynamic configuration and precision control for scalability and reliability.
• Develop software frameworks, operator libraries, and acceleration libraries for NPU platforms to accelerate large‑model AI training.

Required profile

• Currently pursuing a degree in Computer Science, Computer Engineering, AI, robotics or related fields.
• Familiar with large model structures such as Deepseek and Llama and basic optimization of LLM, MoE, and multimodal models.
• Knowledge of AI accelerator hardware (GPU/NPU) and experience optimizing AI systems across software and hardware.

Required skills

• Python programming
• C programming
• C++ programming