模型架构工程师-抖音直播
Core
Design and develop the recommendation system architecture for Douyin live streaming, covering strategy, model, and data architecture to support algorithm iteration and solve system performance/cost/stability challenges.
Role type
Senior IC model architecture engineer (recommendation systems)
Builds
Online inference and offline training pipelines for recommendation models (precision/recall stages) for Douyin live streaming
Domain
Live streaming, recommendation systems, deep learning infrastructure
Deliverable
production ML models
Required skills
GPU hardware architecture understanding, CUDA/CUTLASS/Triton Lang expertise, deep learning framework internals (TensorFlow/PyTorch), challenging problem solving, system performance optimization
Preferred skills
Deep learning operator implementation details, large-scale distributed training experience, MLIR/XLA/TVM familiarity, mathematical analysis of training/inference optimization algorithms