AML 机器学习推理框架研发工程师
Core
Research and development of online AML machine learning inference frameworks to serve ByteDance's various products.
Role type
Senior IC machine-learning inference framework engineer
Builds
High-performance inference services for core business models including computer vision, speech recognition, speech synthesis, and NLP.
Domain
Internet / Machine Learning Infrastructure
Deliverable
production ML models
Required skills
C/C++, Python, Linux, TensorFlow, PyTorch, deep learning models (ResNet, BERT), GPU architecture, CUDA, cuDNN, parallel computing, model optimization (pruning, quantization), TVM, MLIR, XLA
Preferred skills
Deep learning model debugging and tuning, GPU performance analysis, operator implementation details, GAN, reinforcement learning, graph neural networks, AutoML
Responsibilities
Optimize deployment of CV, ASR, TTS, and NLP models; collaborate with algorithm teams to analyze performance bottlenecks and system architecture; implement hardware-software co-optimization for extreme performance.
Seniority
Senior, hands-on IC