AI Infra（异构算力基础设施）研发实习生（J106488）
Core
Develop and build an AI infrastructure platform for heterogeneous computing resources to support large model training and inference workloads.
Role type
AI Infrastructure Research Intern
Builds
Unified access, management, and delivery capabilities for GPU, RDMA, and high-speed network resources.
Domain
AI Infrastructure / Heterogeneous Computing
Deliverable
production ML models
Required skills
Go, Python, C/C++, Linux system development, Distributed systems, Containerization (Docker/Kubernetes), Hardware compatibility testing, Performance optimization, Observability, Automation.
Preferred skills
CUDA, NCCL, GPU monitoring, Multi-node communication, Large model training/inference workflows, HPC.
Responsibilities
Develop GPU/RDMA/network resource management and delivery; Build automated admission and acceptance systems for hardware; Implement full lifecycle resource management; Optimize infrastructure performance and stability; Develop backend services for capacity and cost management; Build observability and automated troubleshooting modules.
Seniority
Intern
