算法研究专家（大模型方向）-国际化内容安全
Core
Researching and developing multimodal large language models (LLMs/VLMs) for video content safety, including pre-training, fine-tuning, and alignment to detect risks and maintain a safe ecosystem.
Role type
Senior IC machine-learning researcher (multimodal LLMs)
Builds
Production multimodal models for risk detection, sorting, and recall in content safety scenarios
Domain
AI/ML, Content Safety, Video
Deliverable
production ML models
Required skills
Multimodal LLM/VLM development, Pre-training and fine-tuning (SFT/RLHF), Few-shot/Zero-shot strategy design, Distributed training optimization, Chain-of-Thought annotation frameworks, Cross-functional technical leadership
Preferred skills
Experience with open-source LLM projects, Publications in top-tier conferences (NeurIPS, ICLR, CVPR, ACL), Deep understanding of content safety risk modeling
Technologies
PyTorch, TensorFlow
