Job Description:
Mission: Leverage hybrid architecture to comprehensively enhance the exchange's risk defense and automated decision-making capabilities, and participate in building the next-generation "AI-Ready" risk control system.
- Black Industry Mining & Graph Analysis: Utilize graph neural networks (GNN) and other algorithms to mine and monitor online fraud intelligence and high-risk address networks by combining massive transaction behaviors and on-chain data.
- Risk Control Strategy Assistance & Risk Prediction: Use AI to assist in risk rule refinement, strategy optimization, and automatic vulnerability detection in existing risk control rules. Develop forward-looking risk prediction models to prevent issues before they occur.
- Market Manipulation & Irregular Trading Detection: Build high-concurrency, low-latency real-time monitoring and interception models for abnormal trading behaviors such as wash trading, spoofing, pump & dump, and front-running.
- Real-Time Anti-Fraud & Hybrid Architecture: Establish and optimize a real-time risk control pipeline combining "traditional machine learning + large model intent recognition" to accurately intercept P2P fraud and abnormal trading groups, reducing financial losses.
- Automated Risk Control Material Review: Build a multimodal review pipeline to achieve automatic material parsing and cross-validation.
Note: Positions available for both AI Risk Control Algorithm Engineer and AI Large Model Infrastructure Algorithm Engineer. Contact via TG for details.
Key Responsibilities:
- Conduct in-depth analysis of transaction patterns and on-chain data to identify fraudulent activities.
- Develop and optimize machine learning models for real-time risk detection and mitigation.
- Collaborate with cross-functional teams to implement risk control strategies and systems.
- Stay updated with the latest trends in financial fraud and adapt models accordingly.
- Ensure compliance with regulatory requirements and industry standards.
Job Requirements:
- Bachelor's or Master's degree in CS, Statistics, Mathematics, or related fields (Master's preferred).
- 5+ years of experience in risk control algorithms.
- High data sensitivity and ability to independently define problems and drive solutions.
- Proficiency in Python and SQL, with experience in large-scale data processing (Hive/Spark).
- Strong ML foundation, familiar with feature engineering and end-to-end model tuning.
- Experience with graph algorithms for fraud detection and group identification.
- Knowledge of sequence models for behavioral anomaly detection.
- Experience with real-time systems (Flink/Kafka) and online inference pipeline design.
- Familiarity with at least one of the following areas: fraud prevention, trading monitoring, P2P anti-fraud, or AML compliance.
Preferred Qualifications:
- Experience in on-chain address analysis and fund tracing (e.g., Chainalysis, TRM).
- Practical experience with LLM in risk control (intent recognition, multimodal review, RAG, workflow orchestration).
- Design experience with risk control rule engines or strategy platforms.
- Understanding of order book mechanisms and market microstructure.
- Background in risk control at top exchanges or large fintech platforms.
- Publications in top conferences like KDD, AAAI, or WWW.
Benefits:
Negotiable
