Fraud and Abuse Analytics Owner

About the role:

You'll own fraud and abuse detection across our cashback, credit, promo, and referral mechanics — building the monitoring that keeps these programs profitable and safe.

What you'll do:

• You'll be the dedicated fraud and abuse analytics owner for loyalty, cashback, credit, and referral programs across Salmon Bank.

• You'll work directly with Risk, Product, and business teams — moving from a suspicious pattern to a tested hypothesis, an investigation, and a permanent monitoring rule.

• The schemes you catch and the controls you build directly shape how much abuse costs the business, and how fast we close each gap.

What you'll own:

• Monitor transactions and customer behavior across cashback, promo, referral, and credit products to catch abuse patterns as they emerge.

• Build and maintain SQL/Python-based reports and monitoring across large transactional datasets.

• Form and test hypotheses on potential abuse schemes, then investigate flagged cases and assess their financial impact.

• Calculate the economics behind each case — potential and actual losses, revenue impact, cost of abuse — to prioritize what gets fixed first.

• Work with business and product teams to close the gaps: propose changes to product terms, promo rules, or cashback mechanics.

• Turn every closed case into a permanent alert or trigger, and revisit past cases on a regular basis to check whether the pattern has returned.

What makes you a strong fit:

• Experience investigating and preventing abuse in banking or financial programs — cashback, credit cards/credit lines, promotions, referral programs, loyalty programs, or quasi-cash/transaction abuse.

• You don't need to have covered every category above — what matters is hands-on experience building continuous monitoring, not one-off analysis.

• Strong SQL and Python (or equivalent scripting), comfortable working directly with large transactional databases.

• A track record of turning a hypothesis into a working monitoring system, not just a one-time report.