Senior Machine Learning Engineer- Identity & Trust

Remitly•Seattle, WA

33d•$196,000 - $245,000•Onsite

About The Position

At Remitly, we believe everyone deserves the freedom to access, move, and manage their money wherever life takes them. Since 2011, we've tirelessly delivered on our promise to customers sending money globally, providing secure, simple, and reliable ways to manage their money, ensuring true peace of mind. Whether it's supporting loved ones back home, growing a business across continents, or pursuing new opportunities abroad, we're not just here to move money— we're here to move our global customers forward. We're looking for builders, reimaginers, and global thinkers who want to work at the intersection of technology, trust, and transformation. If that's you and you're ready to do the most meaningful work of your career—we invite you to join over 2,800 passionate Remitlians worldwide who are united by our vision to transform lives with trusted financial services that transcend borders. About the Role The Trust Machine Learning team protects Remitly's customers by developing intelligent systems that prevent fraud. These systems assess the risk of every transaction and customer interaction, while minimizing the impact on user experience and customer trust. Fraud is one of the hardest applied machine learning problems: an adversary who adapts to every model we ship, extreme class imbalance, labels that arrive late or not at all, and outcomes we only observe for the transactions we approve. Making progress here takes both strong engineering and scientific rigor. As a Senior Machine Learning Engineer, you will develop the models and systems that detect fraud in real time and push the scientific capabilities of our fraud ML program forward. You will frame ambiguous fraud problems as testable hypotheses, advance our modeling techniques beyond established baselines, and carry your ideas from experimentation through to reliable production systems. You will report to the Senior Manager of Trust Machine Learning. This role is based in Remitly's office in Seattle, Washington.

Requirements

A degree in computer science, machine learning, statistics, or a related quantitative field (advanced degree preferred), or equivalent experience
5+ years of experience building and deploying machine learning systems, including a track record of taking novel modeling approaches from idea to production
Deep grounding in ML fundamentals and experimental design: you can reason about why a model works, not just whether it does, and design evaluations that hold up under distribution shift
5+ years of programming experience in Python or equivalent, with hands-on experience in modern ML frameworks (e.g., PyTorch, XGBoost/LightGBM, scikit-learn)
Experience working with cloud platforms (e.g., AWS, GCP, Azure)

Nice To Haves

Experience in fraud, risk, abuse, trust and safety, or another adversarial ML domain
Publications, patents, or open-source research contributions demonstrating applied research impact
Depth in one or more of: graph machine learning, sequence/behavioral modeling, anomaly detection, causal inference, or LLM applications to risk.

Responsibilities

Design, build, and own machine learning models and systems that identify risky transactions and behavior in production
Advance our modeling capabilities with techniques suited to adversarial domains (e.g., graph-based methods for fraud rings, sequence models for behavioral patterns, and semi-supervised or anomaly detection approaches for novel attacks)
Design rigorous offline and online evaluation methodologies that account for extreme class imbalance, delayed and censored labels, selective labeling bias, and adversarial drift
Take modeling work from prototype to production, ensuring models meet real-time latency, reliability, and monitoring requirements
Raise the team's scientific bar through experiment design reviews, model deep-dives, mentoring, and bringing relevant external research into the team's practice
Collaborate with data scientists, risk operations, and business stakeholders to identify emerging fraud patterns and translate them into modeling opportunities