Senior fraud data analyst
Pubblicato il 12-10-2026 - Satispay in Lombardia
Please note that this opportunity requires relocation to our Luxembourg office.
About us
Traditional banking has never really given you control over your financial life. Satispay exists to change that: to empower people through finance . That means making financial services easy and accessible for over 6.5 M users. We started with payments, then benefits, investing, cards... and we're not done. Not even close. We follow the same logic every time: find a real problem, solve it properly, then look for the next one. The destination? To build the most loved financial platform in the world. If you own outcomes rather than tasks, stay close to what users need, and refuse to wait for someone else to fix what's broken — you'll find people here who work exactly in the same way. And together, you'll build something worth being proud of.
What you'll be doing
As our Senior Fraud Data Analyst, you'll handle one of the most critical challenges in digital payments: keeping millions of transactions safe in real time. Here's what that looks like in practice:
Architect Real-Time Strategies – You will be responsible for our real-time decisioning engine, designing and deploying logic that combines ML scores with velocity features to make millisecond decisions.
Precision Rule Development – Translate complex fraud patterns into high-performing rules, optimising the balance between False Positive Rates (FPR) and Gross Fraud Loss. Deconstruct Payment Fraud – Proactively hunt for emerging schemes (ATO, social engineering) and turn these insights into end-to-end mitigation strategies. Optimise the Decisioning Stack – Partner with Product and Engineering to define analytical requirements for new features and third-party data signals.
Calibrate ML & Heuristics – Work with Data Science to operationalise machine learning models, determining how to best "wrap" model scores with heuristic rules. Champion Frictionless UX – Analyse security measures to identify unnecessary hurdles, designing dynamic