Wang Xinzhi

As an enthusiastic AI Systems Researcher, I am deeply invested in exploring and innovating at the forefront of AI technology. My role involves leading and participating in various innovative projects. I not only apply my technical expertise to propel research initiatives but also strive to cultivate a spirit of teamwork and innovation in technology, aiming for breakthroughs in the AI field. My research areas include:

• Training and Inference Acceleration
I am dedicated to optimizing the training and inference processes of deep learning models. This involves developing more efficient algorithms to reduce computational resource consumption and enhance model responsiveness and processing capacity. Through these advancements, my aim is to make AI systems more apt for real-time applications, such as autonomous vehicles and intelligent robotics.
• Efficient Fine-Tuning
I have extensive expertise in efficient fine-tuning, focusing on how to adjust pre-trained models to new tasks or environments with minimal data. This not only improves the adaptability and flexibility of models but also significantly reduces training costs.
• Model Optimization and Resource Management
My work explores achieving optimal model performance with limited hardware resources. This involves algorithm optimization, hardware acceleration techniques, and effective resource allocation strategies.
• Cross-Modal Learning
I have a strong interest in enabling machines to better understand and process multiple types of data (such as text, images, and sound). Research in this area contributes to creating more intelligent and versatile AI applications.
• Evaluation of Large Language Models (LLMs)
A significant strand of my research is dedicated to the rigorous evaluation of large language models. This involves developing comprehensive benchmarks and methodologies to assess the performance, reliability, and ethical implications of LLMs. My work in this area aims to provide critical insights into the strengths and limitations of these models, contributing to their refinement and responsible deployment in various applications.

All these research and development efforts are conducted on NPU/GPU/CPU AI hardware and software systems. Leveraging advanced AI hardware and platform, to provide competitive AI solutions and ecosystem development. I collaborate closely with my colleagues to explore new realms of AI technology, driving innovation in the industry.