Senior Machine Learning Engineer — Fraud Decisioning

SingAI

Hybrid (Amsterdam, NL) · €105K – €128K · Full-time · Posted 6 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

Fraud is one of the few places in our business where everybody notices mistakes immediately. Merchants notice. Risk operations notices. Finance notices. The model can be technically impressive and still be commercially bad if it declines good traffic or lets obvious abuse through. We are looking for a senior engineer who understands that tension and has worked on real-time decision systems before. You will inherit a functioning stack, not a greenfield project, and the job is to make it sharper, faster to debug, and safer to change.

Responsibilities

Improve the fraud decisioning models used across card and wallet transaction flows
Design features and inference patterns that balance predictive power with real-time latency constraints
Own model evaluation, threshold strategy, and post-deployment monitoring for key merchant segments
Work with risk operations and merchant success teams to investigate decision disputes and false-positive patterns
Contribute to technical direction for our broader risk platform

Requirements

6+ years of ML engineering with significant production experience in fraud, risk, ads, or other real-time decision systems
Python for model and service development at production quality
PyTorch or TensorFlow for training and evaluating structured-data models in production
SQL for feature development and investigation across large event datasets
Statistics for experiment design, threshold tuning, and model calibration
Experience deploying low-latency inference systems with clear monitoring and rollback paths