Machine Learning Engineer, Fraud

Whatnot · Consumer · San Francisco, CA · Engineering

Machine Learning Engineer focused on fraud detection and prevention for a livestream shopping platform. The role involves designing, training, and deploying ML models (traditional and LLM-powered), building intelligent user graphs, developing data pipelines and real-time inference systems, conducting data analysis, and implementing model monitoring. Collaboration with cross-functional teams and staying ahead of emerging fraud tactics are key.

What you'd actually do

  1. Design, train, and deploy both traditional ML and LLM-powered models to detect fraudulent behaviors across users, payments, and marketplace interactions.
  2. Lead the end-to-end architecture of fraud detection, prevention, and intervention systems — balancing platform security with a seamless user experience.
  3. Build intelligent user graphs to model behavioral patterns, collusion networks, and account connectivity.
  4. Develop scalable data pipelines and real-time inference systems supporting high-volume, low-latency ML workloads.
  5. Conduct deep behavioral and adversarial data analysis to uncover fraud trends and continuously improve detection accuracy.

Skills

Required

  • Python
  • ML libraries (e.g., scikit-learn, PyTorch, LightGBM)
  • backend development skills
  • deploying ML models to production (batch or real-time)
  • data analysis
  • ETL (SQL, Spark, DBT)
  • data pipeline building
  • fraud detection techniques
  • data orchestration frameworks (Dagster, Kubeflow)
  • feature store design
  • translate business risk into measurable ML solutions

Nice to have

  • LLM-powered models

What the JD emphasized

  • ideally in risk, fraud, or trust & safety domains
  • fraud detection techniques
  • fraud trends

Other signals

  • Develop scalable data pipelines and real-time inference systems supporting high-volume, low-latency ML workloads.
  • Conduct deep behavioral and adversarial data analysis to uncover fraud trends and continuously improve detection accuracy.
  • Partner cross-functionally with Trust & Safety, Payments, and Infrastructure teams to develop features, labels, and model evaluation pipelines.
  • Implement model monitoring and drift detection systems to ensure reliability and responsiveness.
  • Contribute to fraud risk orchestration, combining rules, models, and heuristics for decision automation.