多模态算法专家-抖音风控
Core
Develop and deploy multimodal large-scale models for content understanding and risk control across video, audio, image, and text domains in search, recommendation, and advertising scenarios.
Role type
Senior IC multimodal algorithm expert (risk control)
Builds
Multimodal large models, generative models, and content understanding systems for Douyin, live streaming, and Toutiao
Domain
Internet, risk control, multimodal AI
Deliverable
production ML models
Required skills
deep learning, large language models, multimodal models, generative models, data engineering, model training, mathematical foundations, coding
Preferred skills
open-set content understanding, short video/image algorithms, publications in top-tier conferences (NeurIPS, ICML, CVPR, etc.), competition experience
Technologies
PyTorch, TensorFlow, distributed training frameworks, model evaluation systems
Responsibilities
Research and implement state-of-the-art multimodal model capabilities; build and maintain model training infrastructure and evaluation metrics; track and integrate industry-leading technologies for risk control applications.
Seniority
Senior, hands-on IC
