ML Engineer

The candidate will have responsibilities across the following functions:

Risk Modelling and Business Impact:

• Build and deploy models for:
• Probability of Default (PD)
• Loss Given Default (LGD)
• Exposure at Default (EAD)
• Fraud detection and capture rate optimisation.
• Translate business problems into measurable ML objectives and target variables.
• Drive improvements in risk decisioning, underwriting, and collections strategies.

Machine Learning and Model Development:

• Develop scalable ML models using:
• LightGBM, XGBoost, CatBoost
• Random Forest, CART, Logistic Regression
• Work extensively on tabular datasets (structured financial data)
• Build ensemble and stacking models for improved performance

Feature Engineering and Model Evaluation:

• Perform advanced feature engineering using:
• Weight of Evidence (WoE)
• Information Value (IV)
• Variable Clustering (VarClus)
• Evaluate models using:
• AUC-ROC / Gini coefficient
• F1 Score, Precision, Recall
• Handle class imbalance using:
• SMOTE
• Class weighting
• Threshold tuning

Model Optimisation and Explainability

• Optimise models using:
• Grid Search / Random Search
• Bayesian Optimisation (Optuna preferred)
• Ensure model interpretability using:
• SHAP values
• LIME
• Partial dependence plots
• Communicate model insights effectively to business and risk stakeholders

Data Engineering and Pipeline Development:

• Process large-scale datasets using:
• SQL (advanced level mandatory)
• PySpark / Hive / distributed systems
• Build robust data pipelines for model training and deployment
• Work with large transactional or bureau datasets

Requirements:

• 2 - 5 years of relevant experience in credit risk/fraud analytics.
• Strong hands-on experience with:
• Python (Pandas, Scikit-learn)
• SQL (complex queries, optimisation)
• Expertise in tree-based models (XGBoost/LightGBM).
• Experience with imbalanced datasets in financial use cases.
• Strong understanding of model evaluation metrics beyond accuracy.

Good to Have:

• Experience with:
• PySpark / distributed computing
• Credit bureau / transactional datasets
• Fintech / NBFC / banking domain