AI Infra研发工程师（J99641）
Core
Optimizing training and inference of large language models on domestic GPU hardware and software platforms.
Role type
Senior IC AI Infrastructure Engineer (LLM Optimization)
Builds
Optimized training/inference pipelines and evaluation infrastructure for domestic GPU-based AI models.
Domain
AI Infrastructure / Domestic GPU Computing
Deliverable
production ML models
Required skills
Python, C/C++, PyTorch, Distributed Training, Model Compression, Quantization, GPU Architecture Knowledge, Containerization, Algorithm & Data Structures
Preferred skills
Megatron, vLLM, Zero/Offload, MoE Architecture Tuning, Heterogeneous Computing Acceleration, Large-scale Training/Inference Experience
Technologies
PyTorch, Megatron, vLLM, Domestic GPU Chips
Responsibilities
Optimize LLM training and inference on domestic GPUs; Accelerate software/hardware platforms and operators; Research and implement frontier optimization technologies; Build and maintain evaluation infrastructure for fast/accurate feedback.
Seniority
Mid-Senior, hands-on IC