资深大模型训练推理算子高性能优化工程师-Data AML
Core
Develop high-performance optimization for core operators in multi-modal generation and understanding model training and inference.
Role type
Senior IC machine-learning engineer (model optimization)
Builds
Generative AI models for Douyin, Jianying, and Jimeng products
Domain
AI engineering, multi-modal AIGC
Deliverable
production ML models
Required skills
CUDA, AscendC, BangC, Flash Attention, Conv2d, GEMM, GEMV, low-bit quantization, sparsity optimization, Torch.compile, TVM, MLIR, AIGC model parallel training, inference optimization
Preferred skills
Computer architecture expertise, AIGC algorithm principles
Technologies
CUDA, AscendC, BangC, Torch.compile, TVM, MLIR
