Senior Machine Learning Data Scientist

About the role

You will own the design and execution of advanced machine learning models that power fraud detection and risk decisioning across Extend's post-purchase ecosystem. This role requires you to translate ambiguous business problems into rigorous modeling challenges while balancing accuracy, scalability, and production constraints. You will be responsible for turning high-dimensional transactional and behavioral signals into actionable insights that protect both merchants and consumers. You will act as a technical leader in the Fraud & Machine Learning team, setting standards for experimentation, model evaluation, and rigorous validation. You will work at the intersection of core machine learning research and real-world fraud patterns, ensuring models remain robust against evolving threats. You will communicate complex modeling trade-offs to non-technical stakeholders and influence product and fraud strategy through data-driven recommendations. You will be accountable for the end-to-end integrity of model workflows, from data hypothesis to live system monitoring. You will contribute to building best-in-class machine learning practices that define how Extend protects millions of transactions.

Key facts

What you'll do

• Own the model lifecycle, including requirements gathering, experimentation, model development, evaluation, and creation of model cards, while partnering closely with ML engineers on deployment and production infrastructure.
• Translate complex fraud patterns into well-framed machine learning problems, clearly defining modeling objectives, success metrics, and scenarios where ML adds value over simpler rule-based approaches.
• Design and maintain scalable feature engineering pipelines that support both model development and real-time inference across Extend's transaction streams.
• Monitor model quality in production, tracking performance trends, detecting data drift, and determining optimal retraining strategies to maintain robust fraud detection.
• Partner with leadership, go-to-market teams, fraud operations, product, and engineering to define and execute data-driven fraud strategies aligned with business goals.
• Champion a culture of continuous learning, experimentation, and collaboration across the fraud and broader data science teams to elevate analytical standards.
• Conduct rigorous analysis of user behavior and fraud patterns to inform model features, validation strategies, and detection logic.
• Evaluate emerging machine learning techniques and determine their applicability to Extend's fraud prevention and risk assessment challenges.
• Ensure model outputs are interpretable and actionable for downstream fraud analysts and decision-makers.
• Maintain strong documentation of methodologies, assumptions, and results to enable reproducibility and knowledge sharing.
• Collaborate with data engineering teams to ensure data quality, reliability, and consistency across modeling datasets.
• Drive hypothesis generation, A/B testing design, and measurement frameworks to quantify the business impact of model improvements.

Requirements

• Hands-on, proactive, and analytical professionals who are passionate about using data to solve complex, real-world problems in fraud detection and prevention.
• Bachelor's degree or higher in a quantitative field such as Mathematics, Statistics, Computer Science, Engineering, Operations Research, Physics or related field.
• 3+ years of work experience building and deploying machine learning systems into production in high-stakes environments.
• Strong proficiency in Python for data analysis, modeling, and scripting, as well as SQL for data extraction and transformation.
• Strong understanding of machine learning fundamentals, including model selection, evaluation methodology, feature engineering, and common failure modes such as leakage and overfitting.
• Hands-on experience with PyTorch, scikit-learn, and XGBoost (or similar gradient boosting frameworks) for developing and training models.
• High attention to detail, strong intellectual curiosity, and a deep understanding of user behavior, transaction patterns, and fraud typologies.
• Empathetic, humble, and collaborative team player who communicates clearly with both technical and non-technical stakeholders.
• Candidates must be located within the continental United States due to role and compliance requirements.

Nice to have

• Experience building fraud detection or risk assessment systems at scale, including defining fraud labels and handling label uncertainty.
• Experience with cloud ML platforms, particularly AWS (e.g., SageMaker), and associated MLOps tooling for model deployment and monitoring.
• Experience with graph data and graph-based models, such as those implemented in PyTorch Geometric, for uncovering complex fraud networks.
• Experience with model monitoring and observability tooling (e.g., Arize) to track model performance, data drift, and prediction distributions in production.

Practical notes

• This is a full-time remote position based in the United States.
• Applicants must be eligible to work in the United States without sponsorship for this role.
• The estimated base salary range for this position is $135,000 - $165,000 per year, subject to adjustments based on location, skills, and experience.