广告大模型训练/推理优化研发工程师-Data
Core
Optimizing deep learning engines for training and inference of large-scale models, including compilation, graph fusion, parallel computing, and high-performance operator development.
Role type
Senior IC machine-learning infrastructure engineer (LLM training/inference optimization)
Builds
High-performance operator libraries, distributed training clusters, and optimized inference frameworks for large-scale models
Domain
Artificial Intelligence / Large Language Models / High-Performance Computing
Deliverable
production ML models
Required skills
C/C++/Python, CUDA, Triton, distributed training (FSDP, DeepSpeed), inference frameworks (vLLM, SGLang), SIMD, heterogeneous hardware adaptation
Preferred skills
AscendC, BangC, DCU, next-gen hardware architecture exploration
Responsibilities
Deep optimization of training and inference engines, development of high-performance operator libraries, implementation of distributed training strategies, and research/implementation of frontier technologies
Seniority
Senior, hands-on IC
