Fraud Detection Machine Learning Jobs in Florida

We are looking for a Senior Data Scientist and/or Machine Learning Model Developer to join our Fraud Model Development Team. This role owns end-to-end design, development, validation, deployment partnership, and performance management of high-impact fraud models across SoFi products including Personal Loans, Student Loans, Credit Cards, and Crypto.

The Senior Fraud Model Developer is accountable for measurable fraud loss reduction and false positive improvements within assigned domains. This individual partners closely with Fraud Prevention, Risk, Operations, Finance, Compliance, and ML Platform teams to influence fraud strategy, inform risk tolerance decisions, and ensure models deliver sustained business impact.

This position requires deep expertise in data analytics, statistical modeling, and machine learning, along with the ability to translate complex model performance into clear business outcomes. The ideal candidate brings strong fraud domain knowledge, production ML experience, and a demonstrated ability to manage model risk and complexity at scale

The Senior Fraud Model Developer will help SoFi build and scale high-performing fraud modeling solutions by:

• Owning end-to-end development of fraud models within assigned product or risk domains, from problem framing through production deployment and ongoing monitoring
• Driving measurable reductions in fraud loss, false positives, and operational expenses
• Translating model outputs into business impact metrics and influencing fraud strategy decisions
• Aggregating and synthesizing datasets from multiple data environments to design scalable and reusable modeling frameworks
• Analyzing complex datasets to identify drivers of fraud loss and member friction across products
• Conducting trade-off analysis between fraud loss mitigation, customer experience, and regulatory guardrails
• Establishing and maintaining model monitoring standards (e.g., performance metrics, drift detection, recalibration cadence) to proactively manage model risk
• Investigating external risk data and emerging fraud patterns to inform roadmap prioritization
• Partnering with ML Platform teams to productionize models in AWS and improve lifecycle governance
• Reducing model development cycle time by simplifying processes, improving documentation rigor, and creating reusable components
• Handling escalations related to model performance, risk exposure, or business impact within assigned scope
• Influencing roadmap sequencing and contributing to prioritization discussions based on ROI, regulatory considerations, and level of effort

• 7+ years of advanced quantitative modeling experience, or
• Master's degree and 5+ years of related experience, or
• PhD and 3+ years of related experience, or
• Equivalent practical experience
• Demonstrated ownership of production machine learning models with measurable impact on fraud loss or false positive reduction
• Deep expertise in Python, SQL, and data visualization tools (e.g., Tableau)
• Strong knowledge of statistical methodologies and machine learning techniques (e.g., regression, decision trees, gradient boosting, random forests, neural networks, clustering analysis)
• Experience designing, validating, and monitoring models using metrics such as AUC, KS, precision/recall, and drift detection
• Ability to independently determine modeling approaches, manage trade-offs, and execute solutions with minimal guidance
• Experience partnering cross-functionally with Risk, Fraud Ops, Engineering, Finance, and Compliance stakeholders
• Strong communication skills with the ability to distill complex technical concepts into clear business recommendations
• Demonstrated ability to manage risk exposure within projects and proactively escalate with proposed solutions
  

• Direct experience in fintech, banking, payments, or digital fraud risk
• Familiarity with graph databases and network-based fraud detection
• Experience developing and deploying models in AWS environments
• Experience influencing fraud policy or risk tolerance decisions