AI基础设施架构师/高级AI基础设施研究员-基础设施
Core
Design and evolve end-to-end architecture for ultra-scale, high-availability AI Factory clusters to support efficient pre-training and inference of billion/ trillion-parameter LLMs/VLMs.
Role type
Senior IC AI Infrastructure Architect / Researcher
Builds
Next-generation distributed scheduling systems, high-performance storage, and data pipelines for AI/ML workloads.
Domain
AI Infrastructure, Distributed Systems, High-Performance Computing
Deliverable
production ML models
Required skills
GPU cluster architecture, RDMA/RoCE networking, C++/Go/Rust system programming, distributed training frameworks (Megatron-LM, DeepSpeed), low-latency storage design, performance profiling
Preferred skills
First-principles problem solving, fault-tolerance mechanisms, topology-aware placement, compiler optimization (XLA/TVM)
Technologies
Kubernetes, NCCL, NVLink, InfiniBand, vLLM, Triton
Responsibilities
Design end-to-end AI Factory architecture for 100k+ GPU clusters; Implement distributed scheduling and resource management systems; Optimize full-stack AI/ML compute stack and eliminate system bottlenecks; Design high-throughput distributed storage and ETL pipelines.
Seniority
Senior, hands-on IC