数字人模型工程技术专家-抖音直播（深圳/北京）
Core
Optimizing training and inference algorithms for multi-modal understanding and digital human generation to enhance real-time interaction performance in live streaming.
Role type
Senior IC machine-learning engineer (digital human generation)
Builds
Real-time interactive digital human models for live streaming
Domain
AI, multi-modal systems, large language models
Deliverable
production ML models
Required skills
Python, Go, CUDA programming, Triton, CUTLASS, multi-node/multi-GPU communication, large model training/inference frameworks (Megatron, DeepSpeed, vLLM, TRT-LLM)
Preferred skills
Operator library development
Responsibilities
Optimize multi-modal understanding and digital human generation algorithms; Scale deployment of models to support multi-agent interactions and iterative improvements.
Sourced via bytedance · Listed on CareerPlan, which tracks 853,000+ jobs from 20+ sources.