MLOps Engineer (Assistant Manager) – Credit Risk

About the Role

We are looking for an experienced MLOps Engineer to join our Credit Modelling & Data Management team. This role is responsible for deploying, monitoring and maintaining credit risk scorecard models in production, ensuring models remain stable, reliable and audit-ready. This is not a Data Scientist or Software Developer role. We are looking for someone with hands-on experience in MLOps, model deployment, cloud platforms and production model monitoring.

Principle responsibilities & duties

Data pipeline engineering for scorecard inputs

• Design, build, and maintain scalable pipelines that prepare and refresh scorecard input variables from source systems (LOS, CBS, CCRIS) into DataRobot & AWS Sagemaker, ensuring inputs match the feature definitions used at model development time
• Implement data quality checks, reconciliation routines, and transformation logic so that scoring-time data is provably consistent with development-time data (a common source of silent scorecard degradation).
• Automate ETL/ELT workflows feeding scheduled and on-demand scoring runs, and monitor/troubleshoot pipeline failures or latency issues that could delay or corrupt score generation.

Scorecard deployment & MLOps lifecycle management

• Deploy validated scorecards (DataRobot-hosted and in-house) into production scoring environments, following sign-off from Model Validation.
• Own model versioning and the model registry for all scorecards, maintaining a clear, auditable record of which model version is live, when it was promoted, and what it replaced.
• Support CI/CD for scorecard releases across development, UAT, and production environments, including rollback procedures if a deployed scorecard needs to be reverted.
• Configure and operate DataRobot's native MLOps monitoring (accuracy, data drift, service health) for DataRobot-hosted scorecards.
• Build and maintain custom monitoring for in-house scorecards (e.g. B-Score) that sit outside DataRobot's MLOps tooling, since these have no native platform monitoring.

Scorecard-specific performance & stability monitoring

• Track Population Stability Index (PSI) and Characteristic Stability Index (CSI) on a defined cadence for each live scorecard, flagging breaches against agreed thresholds.
• Monitor score band distribution drift, escalating deterioration patterns to Data Science and business stakeholders.
• Execute and report on champion - challenger comparisons where multiple scorecards or score versions are running in parallel (e.g. B-Score use cases, DataRobot score vs. in-house models).
• Maintain a defined set of retraining/recalibration triggers (e.g. sustained PSI breach, performance decay past threshold) and raise a formal flag to Data Science and Model Validation when a trigger is hit, this role identifies the trigger; it does not decide whether to retrain or redevelop.

Governance, documentation & audit readiness

• Maintain deployment and monitoring documentation to a standard that satisfies Group Audit and regulatory review version history, monitoring logs, incident/rollback records.
• Ensure segregation-of-duties boundaries are respected in practice: this role does not set or change score cutoffs, risk appetite thresholds, or approval strategies, and does not sign off on model validation those remain with Business/Underwriting and Model Validation respectively.
• Support Model Validation and Internal/Group Audit with technical evidence during scorecard reviews and audits.

Stakeholder collaboration

• Work with Data Scientists to understand model logic, feature engineering, and validation findings well enough to operationalise models faithfully and monitor the right things.
• Work with the BI and ITG BA to ensure deployed scores land correctly in upstream origination/decisioning systems and business rules.
• Provide clear, non-technical status updates on scorecard health (deployment status, stability, flagged risks) to the Head of Credit Modelling & Analytics for use in management/committee reporting.

Requirement:

• Bachelor's Degree in Computer Science, Data Analytics, Software Engineering, Information Systems or related discipline.
• Minimum 5 years of experience in MLOps, Machine Learning Platform Engineering or Model Deployment.
• Hands-on experience with DataRobot (or similar AutoML/MLOps platform).
• Experience with AWS (SageMaker, Redshift or equivalent cloud technologies).
• Strong SQL and data pipeline (ETL/ELT) experience.
• Knowledge of CI/CD, model versioning and production deployment.
• Experience in banking, financial services or credit risk analytics will be an added advantage.
• Good communication skills with the ability to explain technical findings to business stakeholders