Job Description:
Mission: Enhance the exchange's risk defense and automated decision-making capabilities through hybrid architecture, 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 analyzing massive transaction behaviors and on-chain data.
- Risk Control Strategy Assistance & Prediction: Leverage AI to assist in risk rule refinement, strategy optimization, and automatic vulnerability detection in existing rules. Develop forward-looking risk prediction models to prevent issues proactively.
- 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: Develop and optimize a real-time risk control pipeline combining traditional machine learning and large-model intent recognition to accurately intercept P2P scams and abnormal trading groups, reducing financial losses.
- Automated Risk Control Material Review: Construct 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 Telegram for details.
Job Requirements:
- Bachelor's or Master's degree in CS, Statistics, Mathematics, or related fields (Master's preferred), with 5+ years of experience in risk control algorithms.
- Strong data sensitivity and ability to independently define problems and drive solutions.
- Proficiency in Python + SQL, and familiarity with large-scale data processing (Hive/Spark).
- Solid ML foundation, with expertise in feature engineering and end-to-end model optimization.
- Familiarity with graph algorithms for black industry mining and group identification.
- Knowledge of sequence models for behavioral anomaly detection.
- Experience with real-time systems (Flink/Kafka) and online inference pipeline design.
- Understanding of business logic and adversarial evolution in at least one of the following areas:
- Black industry countermeasures (fraudulent activities, account farming, batch attacks).
- Trading surveillance (wash trading, spoofing, pump & dump).
- P2P anti-fraud / AML compliance.
- Preferred Qualifications:
- Experience in on-chain address analysis and fund tracing (Chainalysis/TRM).
- Practical experience with LLM in risk control (intent recognition, multimodal review, RAG, workflow orchestration).
- Design experience with risk rule engines or strategy platforms.
- Knowledge of orderbook mechanisms and market microstructure.
- Background in risk control at top-tier exchanges or large fintech platforms.
- Publications in top conferences like KDD, AAAI, or WWW.
Benefits:
Negotiable
