混元大模型基础设施规划和运营工程师/专家（深圳/北京）
Core
Plan and operate hardware infrastructure for the Hunyuan large language model, translating business needs into resource specifications and ensuring efficient resource utilization.
Role type
Senior Infrastructure Planning and Operations Engineer (LLM)
Builds
GPU/CPU clusters, storage systems, and network infrastructure for model training and inference
Domain
Cloud Infrastructure / Large Language Models
Deliverable
infrastructure
Required skills
GPU/CPU/Storage planning, network bandwidth management, data center construction, resource optimization, LLM training requirements analysis, cross-team coordination
Preferred skills
Large-scale cloud resource scheduling, self-built data center experience, complex problem resolution
Technologies
GPU, CPU, Memory, Bandwidth, Storage, Data Center
Responsibilities
Plan hardware resources (GPU/CPU/memory/bandwidth/storage) 6-12 months in advance; Translate LLM business scenarios (pre-training/fine-tuning) into infrastructure metrics; Optimize existing resource allocation and resolve bottlenecks; Participate in data center construction and server selection; Establish mechanisms for cross-team requirement alignment.
