大模型推理优化工程师-抖音
Core
Research and engineering for accelerating inference of AIGC large models (image/video/3D/multimodal) and optimizing GPU resource management during training.
Role type
Senior IC large model inference optimization engineer
Builds
Optimized inference frameworks and deployment services for AIGC models
Domain
AI/ML, Large Language Models, Computer Vision, Graphics
Deliverable
production ML models
Required skills
C++, Python, PyTorch, GPU/NPU operator development, model distillation, quantization, pruning, TensorRT, vLLM, Diffusion model architecture, LLM architecture
Preferred skills
Experience in heterogeneous card adaptation, deep understanding of operator implementation details
Responsibilities
Develop model distillation, quantization, and pruning techniques; adapt models to heterogeneous hardware; optimize GPU utilization during training; deploy engineering services for business use cases; research and implement frontier inference acceleration technologies
