抖音风控基座模型算法工程师（LLM & Agentic AI方向）-Data
Core
Develop and deploy LLMs, MLLMs, and autonomous Agents for risk control in short video, live streaming, e-commerce, and local life scenarios to combat black/grey markets and fraud.
Role type
Senior IC machine-learning engineer (LLM & Agentic AI for risk control)
Builds
Next-generation intelligent risk control infrastructure for 1 billion+ user community ecosystem
Domain
Internet / Risk Control / AI
Deliverable
production ML models
Required skills
Large model training (Pre-training/CPT/SFT/RLHF/RL), Distributed training & inference acceleration, Multi-modal large models (MLLM), Agent system architecture (Tool Use, Planning, Memory, Multi-Agent), Graph-LLM, Behavior sequence modeling, GNN, Model Harness Engineering, Large-scale model service deployment
Preferred skills
Content safety & anti-fraud domain expertise, Top-tier conference publications (NeurIPS/ICML/ICLR/ACL/CVPR), Open source contributions, Cross-team collaboration
Technologies
LLM, MLLM, Graph-LLM, GNN, Agent frameworks
Responsibilities
Lead R&D and optimization of risk control LLMs, MLLMs, and risk knowledge graphs; Design and engineer autonomous risk control Agents for full-link intelligence; Track and integrate frontier AI trends into risk control business scenarios.
Seniority
Senior, hands-on IC
