AI Transaction Velocity Abuse Detector

Gradrix Works

Remote · $160k – $200k / year · Full-time · Posted 5 months ago

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

About the role

Velocity abuse is fraud in slow motion. A single account makes 3 transactions per day — nothing alarming. Zoom out and you see 8,000 accounts following the same pattern, all created within the same 72-hour window, all routing money through the same intermediary. The fraud is visible only at the graph level, and it happens fast enough that by the time a rules engine flags it, the money is gone. We build fraud detection infrastructure for payment networks and neobanks. Our velocity abuse detection system processes 280 million events per day, connecting individual transaction signals to account-level and network-level abuse patterns in real time. The current system catches 78% of velocity fraud rings before they cash out; the target is 90%. You'll redesign the ring detection layer — graph-based features from transaction networks, temporal pattern models that identify coordinated account behavior, and the scoring system that balances detection rate against false positives in a customer-facing context where blocking legitimate users is its own expensive problem.

Fully remote.

Responsibilities

Redesign velocity abuse ring detection using graph-based transaction network features
Build temporal coordination models identifying synchronized account behavior patterns
Design false-positive management framework balancing detection and customer experience
Implement real-time scoring pipeline processing 280M daily events within SLA
Build evaluation framework using confirmed fraud ring case histories
Collaborate with risk operations team on alert investigation and model feedback

Requirements

6+ years in fraud detection or ML engineering with financial crime domain experience
Experience with graph neural networks or graph-based feature engineering
Strong Python, Spark, and SQL skills for large-scale event processing
Deep understanding of precision-recall trade-offs in fraud detection contexts
Track record building real-time ML systems at high transaction volume