data scientist in financial crime

SumUp provides simple and affordable financial tools that help small businesses manage payments, finance, and customer relationships. More than 4 million businesses across 37 markets rely on SumUp as a financial partner.

• Build and operate end-to-end batch training pipelines for transaction-monitoring models;
• Build reliable software for testing, CI/CD, versioning, deployment, monitoring, and rollback across the model lifecycle;
• Improve the maintainability, observability, and scalability of model pipelines;
• Partner with platform and software engineers to make model delivery repeatable and safe;
• Build, maintain, and improve ML models for transaction monitoring;
• Engineer features reflecting AML and Fraud typologies and suspicious behaviours;
• Work with Risk investigators to translate domain knowledge into signals, alerting logic, and calibrated thresholds;
• Analyse AML Risk Score drivers and recommend improvements to features, logic, and thresholds;
• Define and track model and operational metrics, including detection performance, alert volumes, and investigator outcomes;
• Monitor drift and model health, run back-testing, and investigate performance changes;
• Run sensitivity tests on synthetic datasets and assess model behaviour across relevant scenarios and populations;
• Produce model cards, technical documentation, and ML governance artefacts supporting auditability and regulatory review;
• Contribute to system-design documentation and adapt solutions to regional compliance requirements;
• Partner with AML and Fraud Operations, Product, and Engineering to turn ambiguous problems into clear, scalable technical plans;
• Explain trade-offs to technical and non-technical stakeholders;
• Help improve engineering practices, modelling approaches, and understanding of financial-crime risk;
• Share knowledge and support thoughtful experimentation, constructive challenge, and continuous improvement.

• Strong production Python engineering experience, including automated testing, CI/CD, code review, versioning, observability, and operating production services or pipelines;
• Experience deploying and operating ML models in production, including reproducible training, model versioning, deployment, monitoring, incident response, and rollback;
• Hands-on experience with end-to-end ML pipelines from data preparation and training through validation and production use;
• Understanding of appropriate KPIs and evaluation metrics;
• Solid data-engineering fundamentals with complex, multi-source data ecosystems;
• Focus on data quality, lineage, reproducibility, and failure modes;
• Willingness to deepen data-science expertise in modelling, feature engineering, evaluation, and experimentation;
• Clear and confident communication, with the ability to align stakeholders, set expectations, surface risks, and turn ambiguous compliance or operational requirements into concrete technical plans;
• Nice to have: Experience with PySpark or other distributed data-processing technologies, AML, fraud detection, transaction monitoring, or another financial-crime domain, unsupervised learning such as anomaly detection or clustering, feature stores, model registries, alerting-threshold calibration, ML governance artefacts such as model cards, validation reports, or audit documentation, AI systems and tooling.

• Office-first setup from the Berlin office;
• Virtual Stock Option programme;
• Annual L&D budget of €2,000;
• Corporate pension scheme matching up to 20% of contributions;
• 28 Days of paid leave plus public holidays and special leave days;
• Urban Sports Club subsidy, Kita placement assistance, and subsidised office lunches;
• 1-Month sabbatical after 3 years of service;
• Referral bonus.