Job Description:
We are seeking a skilled and motivated System Performance Engineer to join our team. In this role, you will develop and optimize operator libraries for large model inference, training, and fine-tuning processes. You will map operator library implementations to hardware, conduct performance analysis, and identify bottlenecks to maximize system efficiency and performance.

Key Responsibilities:

• Operator Library Development: Develop and optimize libraries for large model inference, training, and fine-tuning, ensuring efficient operation on hardware platforms.
• Performance Analysis: Analyze operator library implementations to identify and resolve performance bottlenecks, optimizing for memory usage, processing time, and data accuracy.
• Optimization Implementation: Engage in the software design and optimization of operator libraries, including instruction-level parallelism, to maximize hardware performance.
• Mapping to Hardware: Map operator library implementation to the underlying hardware, ensuring seamless execution and improved performance.
• Testing and Validation: Perform detailed testing, including memory usage and processing time analysis, to validate optimized library performance.

Qualifications:

• Educational Background: Bachelor's degree or higher in Electronics, Computer Science, or a related field.
• Technical Expertise:
  • Strong understanding of large model algorithms and the principles behind their design.
  • Proficiency in computer architecture and many-core processor principles.
  • Practical experience in CPU performance optimization.
• Programming Skills: Expertise in C/C++ programming for performance optimization.
• Communication Skills: Excellent interpersonal and communication abilities to collaborate effectively within cross-functional teams.

Why Join Us?

• Work on cutting-edge technologies and contribute to the development of high-performance systems.
• Collaborate with a dynamic team of experts in semiconductor and hardware design.
• Make an impact on the performance of next-generation hardware systems.