Staff/Principal Machine Learning Engineer - Bay Area, CA, United States of America

We started a movement in which everyone can win – shoppers, retailers, society and every person on our team. To play fair, trust people and reward them for doing the right thing. We see and feel the impact of our work as more and more people gain financial freedom and retailers g ...

Job description

We started a movement in which everyone can win – shoppers, retailers, society and every person on our team. To play fair, trust people and reward them for doing the right thing. We see and feel the impact of our work as more and more people gain financial freedom and retailers grow across the globe.

Founded 10 years ago in Sydney, Australia, Afterpay has over 24 million active customers globally and more than 250,000 of the best retailers around the world including Anthropologie, Revolve, DSW, GOAT, Finish Line, Levi's, Mac Cosmetics, Ray-Ban and many others.

Afterpay is helping people spend responsibly We empower customers to access the things they want and need, while still allowing them to maintain financial wellness and control, by splitting payments in four, for both online and in-store purchases. Afterpay is deeply committed to delivering positive outcomes for customers. Now under the Block ecosystem, we are focused on supporting our community of shoppers.

We trust in the next generation and share a vision of a more accessible and sustainable world in which people are rewarded for doing the right thing. Join a movement in which everyone can win.

The Role

Fraud and abuse are among the hardest problems in consumer lending. At Afterpay's scale, with millions of active customers and hundreds of thousands of merchant partners, fraud patterns evolve constantly. Adversaries adapt. New vectors emerge across the merchant network, payment flows, and account lifecycle. The modeling problems are genuinely hard: real-time decisioning under adversarial conditions, signal discovery in large and noisy datasets, and building systems that stay calibrated as attack patterns shift.

On the Fraud and Abuse team, you will be a lead individual contributor architecting and building the ML systems that power fraud prevention and abuse mitigation across the lending lifecycle. You will partner closely with Product, Engineering, and Operations to drive platform-level improvements, influence portfolio and risk-appetite decisions, and strengthen the resilience of our lending ecosystem. You will also build agentic features for operations teams, creating AI-native tools that surface emerging fraud vectors, adapt to drift, and enable faster, more precise decision-making at scale.

We use agentic engineering and AI tooling to build reliable, high-velocity workflows that enable this work. That includes code generation, automated testing, documentation, and developer tooling. You will help define how these practices scale across the team in ways that are rigorous, auditable, and trusted.

This is a team that values high output and rigor. We move fast, we test carefully, and we hold our work to a high standard because the systems we build protect real customers and real merchants.

This role is fully remote for candidates based in the US or Canada.

You Will

• Architect and build AI/ML systems that power fraud prevention and abuse mitigation across the Afterpay lending lifecycle.
• Shape lending decision frameworks with advanced analytics, machine learning, and automation to improve precision, adaptability, and speed.
• Analyze large and complex datasets to surface insights that influence underwriting strategy, strengthen ecosystem safeguards, and inform product direction.
• Lead signal discovery and deep experimentation to identify evolving risk patterns and inform the design of next-generation decision systems.
• Build agentic features for operations teams that surface emerging fraud vectors, adapt to changing patterns, and enable faster decision-making at scale.
• Build agentic engineering workflows that accelerate development, testing, and documentation.
• Collaborate cross-functionally with Product, Engineering, and Operations to design systems and features that enhance trust and portfolio health.
• Share modeling context and approaches across teams, helping align how fraud risk is measured, interpreted, and discussed.
• Mentor and elevate fellow scientists and modelers, contributing to modeling standards, best practices, and technical excellence across the team.
• Exercise a high level of autonomy and ownership, driving solutions from problem framing and design through deployment and iteration.

You Have

• A Bachelor's degree in a quantitative field (e.g., Mathematics, Statistics, Physics, Computer Science). Advanced degrees preferred.
• 10+ years of experience in fraud, risk, credit underwriting, or another high-stakes decisioning domain.
• Deep expertise in AI and machine learning methods, statistical modeling, and advanced analytical techniques.
• Strong experimentation skills: you know how to design holdouts, measure lift, and evaluate models beyond aggregate metrics.
• Strong analytical rigor with a deep testing mindset, including experience designing automated validation frameworks for models and decision logic.
• Proficiency with AI-native development workflows. You use LLMs, agentic coding tools, and AI-assisted automation as a regular part of how you build and ship.
• Experience explaining modeling concepts, results, and limitations to senior stakeholders and cross-functional partners.
• Experience working across disciplines in environments with meaningful constraints.

Technologies We Use and Teach

• Python (NumPy, Pandas, scikit-learn, PyTorch, XGBoost, LightGBM)
• AI development tools as core infrastructure: Claude Code, Cursor, Copilot
• MLflow for experiment tracking and model registry
• Internal feature store and model hosting platform
• Prefect and Airflow for orchestration
• SQL / Snowflake
• GitHub
• GCP / AWS

We're working to build a more inclusive economy where our customers have equal access to opportunity, and we strive to live by these same values in building our workplace. Block is an equal opportunity employer evaluating all employees and job applicants without regard to identity or any legally protected class. We will consider qualified applicants with arrest or conviction records for employment in accordance with state and local laws and "fair chance" ordinances.

We believe in being fair, and are committed to an inclusive interview experience, including providing reasonable accommodations to disabled applicants throughout the recruitment process. We encourage applicants to share any needed accommodations with their recruiter, who will treat these requests as confidentially as possible. Want to learn more about what we're doing to build a workplace that is fair and square? Check out our I+D page.

While there is no specific deadline to apply for this role, U.S. roles are typically open for an average of 55 days before being filled by a successful candidate. Please refer to the date listed at the top of this job page for when this role was first posted.