推荐算法工程师（多模态生成式推荐方向）-抖音内容理解
Core
Build multi-modal representation systems for video/live streaming to fuse visual, audio, and text features, optimizing user interest profiling and content matching accuracy.
Role type
Senior IC multi-modal generative recommendation engineer
Builds
Recommendation systems for short video and live streaming platforms
Domain
Entertainment / Video Streaming / Multi-modal AI
Deliverable
production ML models
Required skills
Multi-modal representation learning, Generative AI, Large Language Models, Deep Learning, Recommendation System Architecture, Reinforcement Learning, Collaborative Filtering, Matrix Factorization, Data Analysis
Preferred skills
Publications in top-tier conferences (NIPS, ICML, CVPR, KDD, SIGIR, etc.), Experience with generative recommendation paradigms
Technologies
PyTorch, TensorFlow, Spark, Hadoop, Kubernetes
Responsibilities
Design and optimize full-link recommendation algorithms (recall, coarse ranking, fine ranking, re-ranking), Build offline evaluation systems aligned with online business goals, Engineer generative recommendation models for live streaming scenarios, Collaborate with product and engineering teams to deliver value
Seniority
Senior, hands-on IC