电商场景LLM/VLM/AIGC训练工程师-TikTok Shop
Core
Research and implement optimization algorithms for e-commerce LLM/VLM/AIGC training to enhance speed and efficiency.
Role type
Senior IC machine-learning engineer (training optimization)
Builds
Production ML models for e-commerce product lines
Domain
E-commerce, Large Language Models, Computer Vision
Deliverable
production ML models
Required skills
C/C++ or Python, Linux, deep learning fundamentals, neural network architectures, AutoML, model training optimization, TensorFlow/PyTorch/MxNet, distributed training frameworks (data/model/pipeline parallelism), LLM/VLM/AIGC algorithms, FSDP/DeepSpeed/Megatron
Preferred skills
Experience with e-commerce LLM/VLM/AIGC training acceleration, hands-on experience with FSDP/DeepSpeed/Megatron
Technologies
CUDA, TensorFlow, PyTorch, MxNet, FSDP, DeepSpeed, Megatron
Responsibilities
Optimize training speed and efficiency using distributed training frameworks; solve engineering problems across CT, SFT, and RL stages; collaborate with algorithm teams to analyze performance bottlenecks; deploy and optimize models for e-commerce products; develop AI toolchains and technical ecosystems.
Seniority
Mid-level, hands-on IC
