Senior Fraud Data Scientist - Production ML & Card Risk

Virtusa Middle East FZ-LLC; Dubai; On-site; AED 300,000 - 460,000; Full time

Job summary

Virtusa Middle East FZ-LLC invites an experienced data science leader to develop and optimize card fraud detection models for issuing and acquiring portfolios in a fast-paced banking environment. The role emphasizes deploying models with SAS SFD, building robust MLOps pipelines, and rigorous validation, with collaboration across fraud risk, IT, and data teams to deliver production-ready solutions.

Candidates should have 7+ years in data science, with at least 3 years in card fraud modeling, and

Qualifications

• Master’s or Ph.D. in Data Science, Statistics, Computer Science, or related field.
• 7+ years in data science roles, with at least 3 years in card fraud modeling.
• Proven track record of deploying models in production environments.

Responsibilities

• Card Fraud Model Development: Develop and optimize card fraud detection models for issuing and acquiring portfolios.
• Model Hosting: Deploy models using SAS SFD and integrate with production systems.
• MLOps: Establish and maintain MLOps pipelines for CI/CD and monitoring of fraud models.
• Evaluation: Conduct rigorous validation, stress testing, and performance benchmarking with fraud ops teams.
• Collaboration: Partner with fraud risk, data engineers, IT, and business teams; communicate insights to senior management.

Skills

SAS

Python

SQL

TensorFlow

PyTorch

Scikit-learn

MLflow

Kubeflow

CI/CD

Fraud modeling

Education

Master's or PhD in Data Science

Tools

MLflow

Kubeflow

CI/CD pipelines

Job description

Virtusa Middle East FZ-LLC invites an experienced data science leader to develop and optimize card fraud detection models for issuing and acquiring portfolios in a fast-paced banking environment. The role emphasizes deploying models with SAS SFD, building robust MLOps pipelines, and rigorous validation, with collaboration across fraud risk, IT, and data teams to deliver production-ready solutions.

Candidates should have 7+ years in data science, with at least 3 years in card fraud modeling, and