星星海实验室- AI Infra CPU 研发工程师
Core
Analyze and optimize CPU performance for AI agent scenarios, including reinforcement learning and tool use, by building benchmarking systems and driving hardware-software co-optimization.
Role type
Senior IC CPU infrastructure engineer (AI workloads)
Builds
Production-grade CPU performance benchmarks and optimization frameworks for AI agents
Domain
AI infrastructure / High-performance computing
Deliverable
production ML models
Required skills
Linux performance analysis, CPU microarchitecture knowledge, Agent/RL system understanding, hardware-software co-optimization, profiling and tracing, NUMA and memory bandwidth management
Preferred skills
Experience with distributed systems, PCIe/CXL protocols, SMT and cache optimization
Technologies
Linux, CPU performance counters, profiling tools, tracing tools
Responsibilities
Analyze CPU load and select optimal configurations for agent scenarios; Build end-to-end performance analysis and benchmarking systems for AI tasks; Identify and resolve bottlenecks in CPU, cache, memory, NUMA, I/O, and scheduling
