大模型训练加速工程师-智能创作（上海/深圳）
Core
Optimize visual large model training pipelines to enhance efficiency, speed, and scalability.
Role type
Senior IC large model training acceleration engineer
Builds
Optimized training pipelines and distributed training strategies for visual large models
Domain
AI/ML, Large Language Models, Computer Vision
Deliverable
production ML models
Required skills
Python, C++, PyTorch, distributed training strategies, performance profiling, GPU architecture understanding
Preferred skills
Megatron, Deepspeed, CUDA programming, tensor parallelism, sequence parallelism, pipeline parallelism, operator-level optimization
Sourced via bytedance · Listed on CareerPlan, which tracks 854,000+ jobs from 20+ sources.