Staff Applied Machine Learning Engineer - Fraud & Abuse

As a Staff Applied Machine Learning Engineer focused on Fraud & Abuse, you will design, build, and operate production ML decision systems that reduce payment fraud, account takeover, identity abuse, merchant and marketplace risk, scams, and other adversarial activity across Block.

The team optimizes for reliable decisions, safe deployment, and measurable customer outcomes - preserving access for good customers while reducing fraudulent, abusive, or unsafe activity.

You should be comfortable owning production systems end to end: data contracts, low-latency inference, batch scoring, feature quality, online/offline consistency, model deployment, monitoring, incident response, rollback, and outcome feedback loops. The work combines large-scale ML decisioning with AI-assisted operations: surfacing evidence, simulating controls, accelerating triage, and improving feedback loops while preserving human judgment in high-stakes decisions.

You will work closely with ML modelers, product engineers, risk analysts, compliance partners, and operations teams to respond quickly to evolving abuse patterns without creating unnecessary friction or harm for legitimate customers.

You Will

• Build and operate real-time and batch ML decisioning systems for payment fraud, scams, identity and account integrity, merchant and marketplace risk, and abuse prevention.
• Integrate behavioral, graph, device, network, event-stream, and third-party signals into low-latency model serving, decision APIs, and product controls.
• Own the production lifecycle for risk decisions, including data contracts, feature quality, online/offline consistency, monitoring, drift detection, safe rollout, rollback, and incident response.
• Develop feedback loops and verified AI-assisted workflows for triage, investigation support, alert clustering, graph exploration, simulation, and post-incident learning.
• Partner with modelers, analysts, product, compliance, and operations to balance fraud losses, customer access, false positives, product velocity, support burden, and long-term trust.
• Create reusable decision and evaluation capabilities that product services, internal tools, and AI-assisted workflows can safely consume., We may use automated AI tools to evaluate job applications for efficiency and consistency. These tools comply with local regulations, including bias audits, and we handle all personal data in accordance with state and local privacy laws.

• 12+ years building and operating production software and ML systems for business-critical products.
• Deep expertise in fraud/risk domains such as payment fraud, identity/account integrity, merchant or marketplace risk, scams, trust & safety, abuse prevention, or compliance decisioning.
• Strong production ML judgment across feature pipelines, model serving, evaluation, monitoring, low-latency integration, safe rollout, and incident response.
• Sound judgment around false-positive tradeoffs, noisy labels, adversarial behavior, customer harm, and cross-functional decisions.
• Experience using AI-assisted engineering tools with appropriate verification, testing, and review for high-stakes systems.

Nice to Have

• Experience with graph-based fraud detection, behavioral sequence models, embeddings, entity resolution, anomaly detection, or human-in-the-loop review.
• Experience building fraud operations tooling for triage, case management, alert clustering, graph exploration, or policy simulation.
• Experience with regulated financial services, model governance, auditability, explainability, or decision logging., We do not expect candidates to have used our exact stack. We do expect strong production engineering fundamentals, deep domain expertise in intelligent ML systems, and judgment about how ML-derived signals should be used safely in customer-impacting products. Examples of technologies and methods include:
• Python, Java, Kotlin, SQL.
• TensorFlow, PyTorch, XGBoost/LightGBM, embeddings, deep learning, and tree-based modeling ecosystems.
• Kafka or other event-streaming systems, batch data pipelines, feature stores, workflow orchestration, and model-serving systems.
• Cloud infrastructure, Kubernetes, data warehouses/lakehouses, monitoring, observability, coding agents, evaluation harnesses, and agent-assisted operations tooling.

Every benefit we offer is designed with one goal: empowering you to do the best work of your career while building the life you want. Remote work, medical insurance, flexible time off, retirement savings plans, and modern family planning are just some of our offering. _Check out our other benefits at Block.