Graduate GPU AI Platform Engineer — System Optimization

ByteDance; San Jose (CA); On-site; USD 162,000 - 317,000; Full time

Job summary

ByteDance is seeking a graduate-level GPU/AI Platform Architect to optimize giant model systems, including LLM workloads. You will work on architecture, validation, and performance tuning for high-performance server and storage platforms at scale in a cutting-edge hardware environment.

The role emphasizes deep system-level optimization, collaboration with industry groups, and hands-on development in C/C++ and Python, with opportunities to shape next-generation AI infrastructure.

Qualifications

• PhD in Electrical, Computer Engineering, Computer Science or related field.
• Thesis on GPU/AI platform architecture or co-design.
• Deep understanding of GPU/AI SoC, interconnects and memory subsystems.
• Experience in GPU/AI system performance optimization or HW/SW co-design.
• Strong C/C++ and Python development skills.
• Knowledge of GPU/AI virtualization, DL architectures and distributed systems.

Responsibilities

• Develop application benchmarks, tools and optimization methods for GPU/AI systems including giant model training/inference.
• Identify bottlenecks via data-driven study and drive HW/SW co-design improvements.
• Develop GPU/AI system TCO model from benchmarks and performance data.
• Collaborate with industry groups to explore new standards and contribute research results.
• Coordinate with partners to set up POC or prototypes to test new designs.

Skills

C/C++

Python

GPU/AI architecture

System performance optimization

Hardware-software co-design

Education

PhD in EE/CE/CS

Tools

Git

Job description

ByteDance is seeking a graduate-level GPU/AI Platform Architect to optimize giant model systems, including LLM workloads. You will work on architecture, validation, and performance tuning for high-performance server and storage platforms at scale in a cutting-edge hardware environment.

The role emphasizes deep system-level optimization, collaboration with industry groups, and hands-on development in C/C++ and Python, with opportunities to shape next-generation AI infrastructure.