AI Engineer - Anti-fraud

About UsSiFi is a fast-growing B2B fintech transforming corporate expense management in Saudi Arabia. As a licensed Electronic Money Institution regulated by the Saudi Central Bank (SAMA), we issue cards, move money, and manage spend for thousands of Saudi businesses.The RoleWe're hiring an AI Engineer to work across our risk and financial crime domain — one of the most data-rich and adversarial parts of the business.You won't be handed a narrow spec. You'll dig into open-ended problems, work out where our current systems and partner platforms fall short, and build the models and tooling that close the gap. We care far more about how you think than what you've worked on before.What You'll Do Take open-ended problems, frame them, find the data, and build something useful Build models and analytics on top of our existing platforms — anomaly detection, behavioral profiling, alert prioritization Measure how well our systems and partner platforms actually perform, and use that to shape how we configure and improve them Work with our technology partners as a technical counterpart, pressure-testing their outputs Build internal tooling that makes our risk and operations teams faster Apply modern AI, including LLM-based tooling, to workflows that are still manual Keep everything explainable and auditable to regulatory standard RequirementsWhat We're Looking ForRequired 2-3 years building data or ML systems that made it to production Strong Python and SQL, and comfort with large, messy, real-world data A real problem solver — curious, resourceful, and drawn to questions nobody has answered yet You know the difference between a model that scores well offline and one that works in production Confident holding your own in a technical review with an external partner Saudi nationals only Nice to Have Exposure to risk, fraud, AML, cybersecurity, or another adversarial domain — not required, we'll teach you Real-time or streaming data Graph or network analysis Familiarity with payments infrastructure Awareness of SAMA expectations around data and model governance