About the role
We are seeking an experienced Data Scientist to lead the development of ML-driven applications and behavioural scorecards for retail lending portfolios. The role requires strong statistical rigour, end-to-end model ownership, and the ability to translate credit risk appetite into scalable data-driven strategies.
Responsibilities:
- Design and deploy statistical models, including but not limited to application and behavioural scorecards, using advanced ML techniques.
- Lead end-to-end lifecycle: problem framing, data extraction, feature engineering, model development, validation, deployment, and monitoring.
- Develop statistically robust models using techniques such as (but not limited to): Gradient Boosting (XGBoost, LightGBM, CatBoost), Random Forest, ensemble methods, and neural networks (MLP).
- Apply rigorous evaluation frameworks (AUC, KS, Gini, Lift, PSI, stability analysis).
- Address imbalanced datasets, reject inference, segmentation, and portfolio drift.
- Build challenger models and drive continuous performance optimisation.
- Develop and refine credit underwriting strategies aligned to risk appetite, approval rate targets, and portfolio profitability.
- Conduct portfolio analytics and recommend cut-off, limit, and pricing strategies.
- Collaborate with Risk, Policy, Product, and Tech teams to translate business requirements into deployable decisioning frameworks.
- Ensure robust monitoring, back-testing, and periodic recalibration.
Requirements:
- Bachelor's in engineering, business, analytics, data science, or a related field.
Technical Requirements:
- Strong foundation in statistics, hypothesis testing, sampling theory, and classification methodologies.
- Proven experience in ML-based credit modelling, model governance and monitoring.
- Expertise in handling large structured financial datasets.
- Proficiency in Python, PySpark, Scikit-learn, and boosting frameworks.
- Experience building scalable training pipelines and supporting production deployment.