高级风控算法专家（本地生活方向）-TikTok业务风控
Core
Design and build end-to-end risk control algorithm systems for TikTok's local life business to protect users, merchants, and transactions from fraud and abuse.
Role type
Senior IC machine-learning engineer (local life risk control)
Builds
High-precision, low-latency risk identification and scoring models for merchant onboarding, store claiming, false group buying, transaction fraud, and刷单 (brushing)
Domain
E-commerce local life + financial risk control
Deliverable
production ML models
Required skills
Machine learning theory, Deep learning, Classification, Ranking, Sequence modeling, Anomaly detection, XGBoost, DeepFM, Transformer, Graph Neural Networks, Python/C++/Go, PyTorch/TensorFlow, Distributed systems, Offline training + Online inference pipelines
Preferred skills
Agentic AI, LLM fine-tuning (SFT/RLHF), Multimodal representation learning, Geospatial risk control, E-commerce local life fraud experience, POC to production deployment
Technologies
XGBoost, DeepFM, Transformer, Graph Neural Networks, PyTorch, TensorFlow, Python, C++, Go
Responsibilities
Design full-link risk control algorithms for merchant onboarding and transaction fraud; Build high-precision models using multimodal data; Establish black-market countermeasures and emergency response systems; Lead technical research and POC validation for Agentic AI and large models; Coordinate cross-team collaboration for architecture and deployment
