Improve fraud detection accuracy and signal quality across Sumsub products using live production data.

Responsibilities:

  • Investigate emerging fraud patterns, anomalous activity, and gaps in detection logic.
  • Develop and tune fraud detection approaches using rules, supervised and unsupervised models, anomaly detection, and link-based analysis.
  • Evaluate fraud signals, analyse client traffic and alert flows, measure false positives and missed fraud, and recommend improvements.
  • Build and optimise baseline rule sets and proactive alerts for common and emerging fraud scenarios.
  • Turn analytical findings into product recommendations and support complex, data-intensive client cases.

Requirements:

  • 4+ years of experience in anti-fraud, risk analytics, applied data science, or a related analytical role.
  • Strong knowledge of exploratory data analysis and classical machine learning, including supervised and unsupervised approaches.
  • Practical experience with anomaly detection, clustering, graph or link-based analysis, and fraud prevention or analytics.
  • Strong Python and SQL skills; experience with ClickHouse, Trino, Superset, dbt, DataHub, or comparable analytics tools.
  • Ability to assess data quality, bias, leakage, overfitting, model results, false positives, and false negatives.
  • Clear communication of complex findings and fluent English.

Conditions and benefits:

  • Remote-first work with flexible hours and no mandatory office days.
  • Birthday holiday, 10 personal days annually, seven sick days without paperwork, and additional Christmas and New Year time off.
  • Market-benchmarked compensation, professional development support, covered team offsites, and necessary tools and hardware.
  • Hiring stages: TA screening, Hiring Manager Interview, and Final Interview.