Fraud Data Scientist (Fintech)

Location: Berlin, Germany; hybrid work with up to 3 days per week from home

Company

Provides B2B Buy Now, Pay Later payment methods and digital checkout solutions based on proprietary machine-learning-supported risk models and a scalable technology platform.

What you will do

• Design, build, test, and ship production-ready anti-fraud machine learning models.
• Model debtor behavior, identify risk factors, and optimize real-time transaction decisioning.
• Balance precision and recall under severe class imbalance, weighing customer friction against financial loss.
• Monitor deployed models for drift and adversarial adaptation, then retrain or recalibrate them as fraud patterns evolve.
• Own ML deployment and operationalization for real-time, low-latency services, collaborating on containerization and event-driven architectures.
• Work with data and software engineers, analysts, and product managers while communicating findings to technical and non-technical stakeholders.

Requirements

• 3–5+ years of experience in a quantitative or machine learning role, ideally in fintech or a high-transaction environment.
• Advanced proficiency in Python, including pandas, scikit-learn, and xgboost, plus SQL with Snowflake, Postgres, or MySQL.
• Deep expertise in classification models, anomaly detection, and graph-based methods such as graph neural networks and entity-link analysis.
• Hands-on experience building and deploying production ML models, including real-time API serving for online ML systems.
• Strong MLOps knowledge, including Docker, Kubernetes, and event-driven architectures.
• Ability to solve complex problems, manage stakeholders, and translate technical findings into business recommendations.

Nice to have

• Experience with Metaflow, Apache Flink, or similar ML orchestration and MLOps tooling.
• Experience implementing LLM-based workflows such as agentic pipelines, retrieval-augmented generation, or LLM-assisted feature extraction for fraud detection or risk signals.

Culture & Benefits

• Flexible working hours and a hybrid work model.
• 30 days of vacation per year, sabbatical opportunities, and additional child sickness leave for parents.
• Virtual Shares Incentive Program and yearly development budget.
• Discounted Berlin public transport, Deutschland-Ticket, or JobRad access.
• Free German group classes and an English-speaking, multicultural environment with more than 40 nationalities.
• Company and team events, interest groups, a run club, and game nights.