Senior Data Scientist – Credit Risk

Risk Team | Cash Credit

Own the models. Challenge the status quo. Improve credit decisions every day.

About the role

We are looking for a hands-on Senior Data Scientist to take end-to-end ownership of our consumer credit risk models and help us raise the bar on how scoring is developed, monitored and governed.

This is not a role for someone who wants to build a model, hand it over and move on. We are looking for a senior individual contributor who will treat the scoring ecosystem as a product: know how it behaves in production, watch the key signals, spot deterioration or opportunity early, and drive improvements from analysis through implementation.

You will work at the intersection of Data Science, Credit Risk and business decisioning, with direct visibility of how your work affects approvals, portfolio quality, automation and profitability. Experience with scoring models for individuals in consumer lending, retail banking or a comparable high-volume credit environment is strongly preferred.

What you will own

Credit Risk Modelling & Decisioning

·   Own the full model lifecycle for application scorecards, behavioural scorecards, risk segmentation and affordability models: development, calibration, implementation, monitoring and redevelopment.

·   Build and enhance models for individuals using both traditional scorecard methods and modern machine-learning techniques; know when to use each and how to explain the trade-offs.

·   Develop and test challenger models using approaches such as CatBoost and Gradient Boosting, and maintain a pragmatic champion/challenger framework.

·   Translate model outputs into actionable credit strategy: cut-offs, risk bands, routing, automation rules and segmentation logic.

·   Continuously look for opportunities to improve the existing scoring setup rather than waiting for a formal redevelopment cycle.

Model Performance – every day, not once a quarter

·   Keep a close eye on production KPIs and model behaviour, with a strong bias for action when something moves unexpectedly.

·   Monitor Gini, ROC/AUC, KS, PSI, CSI, calibration, score distributions, approval rates, overrides and other key diagnostics.

·   Connect model metrics to real portfolio outcomes: FPD, early delinquency, vintage performance, bad rates, approval mix and performance by score band, channel and customer type.

·   Define clear thresholds and early-warning signals for drift, instability and material performance deterioration.

·   Investigate root causes, quantify impact and propose the right response: recalibration, policy adjustment, feature review or full model redevelopment.

Model Governance & Validation

·   Own model governance standards across the scoring landscape, including model inventory, ownership, versions, changes, approvals, limitations and monitoring thresholds.

·   Perform robust back-testing, stability testing, calibration analysis and independent challenge of assumptions and model behaviour.

·   Lead analytical preparation for annual validation and coordinate with external validators, auditors and internal governance forums.

·   Maintain clear, decision-useful documentation covering methodology, data, assumptions, limitations, validation results and monitoring conclusions.

·   Present model performance, issues and recommendations to the Risk Committee and senior stakeholders in a clear and commercially relevant way.

IFRS 9 & Portfolio Analytics

·   Build, maintain and enhance PD, LGD and EAD models used in IFRS 9 Expected Credit Loss calculations.

·   Maintain and challenge the IFRS 9 ECL framework, including staging logic, calibration and macroeconomic overlays.

·   Support monthly and quarterly provisioning with accurate, reproducible and well-controlled analytics.

·   Explain movements in provisions and distinguish portfolio effects, model effects and assumption changes.

Data, Systems & Production

·   Work closely with Data Engineering, IT and Risk to improve data quality, pipelines and model inputs.

·   Build reliable and reproducible analytical workflows using Python and SQL.

·   Take models all the way into production: specifications, testing, UAT, implementation checks and post-deployment monitoring.

·   Make sure what is live in production matches the approved methodology and investigate input/output anomalies quickly.

·   Help build the monitoring datasets, dashboards and automated alerts needed for scalable model governance.

What we are looking for

Experience & technical depth

·   Typically 5+ years of relevant experience in credit risk modelling, Data Science or advanced analytics, with meaningful hands-on exposure to consumer / retail credit risk.

·   Proven experience building and owning scoring models for individuals, ideally application and/or behavioural scorecards, and following them after go-live.

·   Strong track record in model monitoring, recalibration, validation and governance – not just model development.

·   Strong Python and SQL skills and the ability to independently work through large, imperfect real-world datasets.

·   Hands-on experience with tree-based models, boosting algorithms, sampling techniques and machine-learning methods.

·   Strong practical knowledge of logistic-regression scorecards, including WOE/IV, binning, feature selection, calibration and score scaling.

·   Strong understanding of Gini, ROC/AUC, KS, PSI, CSI, calibration and related model diagnostics.

·   Good understanding of consumer credit portfolio KPIs and the ability to connect model behaviour to approval, delinquency, vintage and profitability outcomes.

·   Knowledge and preferably practical experience with PD, LGD and EAD models.

The mindset we value

·   Ownership mindset: you own outcomes, not just analytical outputs.

·   Proactive and curious: you spot signals in the data, ask the next question and push until the answer is clear.

·   Builder mentality: you improve the existing setup, introduce better controls and leave systems stronger than you found them.

·   Commercially minded: you understand that the best model is the one that improves real credit decisions, not simply the one with the highest offline metric.

·   Comfortable in the details and in the room: able to investigate code and data deeply, then explain the conclusion clearly to senior stakeholders.

·   Constructive challenger: willing to question established assumptions and support recommendations with evidence.

·   High standards for data quality, reproducibility, documentation and governance.

·   Good command of English.

Education

·   Bachelor’s or Master’s degree in Statistics, Mathematics, Data Science, Econometrics, Finance, Physics or a related quantitative field.

What you can expect from us

Real ownership

A senior role with genuine responsibility for the scoring and model-governance setup, not a narrow modelling workstream.

Visible impact

Your work will directly influence credit decisions, approval strategy, portfolio quality and automation.

Room to build

You will have the mandate to challenge, improve and modernise models, monitoring and governance where there is a clear business case.

Senior exposure

Close interaction with Risk leadership and regular visibility with senior management and governance forums.

A pragmatic environment

We value rigorous analysis, but also speed, judgement and the ability to turn insight into something that works in production.

Recognition for contribution

Expertise, initiative and measurable impact matter more than hierarchy.

If this sounds like you

If you are an experienced credit-risk modeller who wants to own the full lifecycle, stay close to production performance and continuously improve how scoring works in the real world, we would like to hear from you.

Send us your CV and let’s talk.