Machine Learning Engineer (LLM)- E-commerce Risk Control
Core
Develop and deploy machine learning models and LLM-based agentic workflows to detect and prevent fraud, abuse, and anomalies in global e-commerce transactions.
Role type
Senior IC machine learning engineer (LLM & risk control)
Builds
Real-time risk detection systems, agentic decision-making pipelines, and fraud monitoring tools for TikTok Shop and Toko.
Domain
E-commerce, fraud detection, large-scale real-time systems
Deliverable
production ML models
Required skills
Python, Java/C++, supervised/unsupervised learning, LLM post-training, agent-based systems, RAG, prompt engineering, knowledge distillation, BERT architectures, large-scale behavioral data analysis
Preferred skills
PhD in ML/NLP, LangChain, explainable ML (SHAP/LIME), graph neural networks, e-commerce fraud background, LLM integration in decision systems
Technologies
LangChain, RAG, BERT, Transformers, SHAP, LIME, Graph Neural Networks
Responsibilities
Design prompt engineering and reasoning workflows for LLM-based decisions; build agentic workflows for complex risk cases; work with large-scale datasets to uncover fraud signals; collaborate with ops and product teams to implement scalable risk strategies.
Seniority
Senior, hands-on IC
