Staff Software Engineer - Real-Time Payments Fraud & Risk

Anthropic Limited; New York (NY); Hybrid; USD 320,000 - 485,000; Full time

Job summary

Anthropic is seeking a software engineer for its Fraud Prevention team to design and build real-time risk decisioning systems and to manage the dispute and chargeback lifecycle across subscriptions, trials, and promotions.

The role emphasizes collaboration with finance, support, legal, and data science, plus external processors and platform partners, with a focus on scalable, hybrid rules-and-ML risk approaches and ownership of key risk metrics.

Qualifications

• Proficiency in Python and SQL; experience with data analysis tools.
• Experience building or operating fraud/risk detection systems in production.
• Strong communication skills to explain complex tradeoffs to non-technical stakeholders.

Responsibilities

• Design and build real-time risk decisioning that scores transactions at authorization time.
• Build tooling for the dispute and chargeback lifecycle from review queues to evidence collection.
• Engineer fraud signals at scale—device fingerprinting, BIN and issuer signals, velocity features, cross-account linkage.
• Own metrics such as loss rate, dispute rate, authorization impact, false positives.
• Lead investigations into emerging fraud patterns with multi-layered defenses.
• Collaborate cross-functionally with finance, support, legal, data science, and external processors.

Skills

Python

SQL

Data analysis tools

Communication skills

Education

Bachelor's degree

Tools

Stripe

Adyen

In-app purchases

Chargeback tooling

Job description

Anthropic is seeking a software engineer for its Fraud Prevention team to design and build real-time risk decisioning systems and to manage the dispute and chargeback lifecycle across subscriptions, trials, and promotions.

The role emphasizes collaboration with finance, support, legal, and data science, plus external processors and platform partners, with a focus on scalable, hybrid rules-and-ML risk approaches and ownership of key risk metrics.