Credit Risk Data Scientist

Company Description LendGo is a digital micro-lending and financial technology company dedicated to expanding access to responsible credit across underserved and emerging markets in Africa. The company delivers fast, transparent, and affordable financial solutions through a fully digital lending experience.
Location: South Africa
Experience: 3–5+ years in credit risk analytics, predictive modelling or consumer lending decisioning
Role Purpose
LendGo is seeking a hands-on Credit Risk Data Scientist to develop and continuously improve data-driven credit decisioning models that identify creditworthy borrowers, predict repayment behaviour, reduce defaults and support profitable loan book growth.
The successful candidate will use customer transactional data, bank statements, credit bureau information and historical repayment performance to build predictive credit risk scorecards and improve lending decisions.
Key Responsibilities
1. Credit Risk Modelling & Predictive Analytics
Develop machine learning models to predict customer repayment probability and first-payment defaults.
Build and validate application credit risk scorecards using historical customer data.
Identify transactional patterns that distinguish reliable borrowers from high-risk customers.
Apply statistical modelling, logistic regression, decision trees and gradient boosting techniques.
Continuously monitor model accuracy and recalibrate decision thresholds.
2. Bank Statement & Transactional Data Analysis
Analyse three months of customer bank transactions to identify credit risk indicators.
Evaluate salary consistency, spending behaviour, disposable income and cash flow stability.
Identify patterns involving gambling, loan stacking, failed debits, salary depletion and recurring financial commitments.
Develop predictive variables from transactional banking data.
3. Credit Bureau & Repayment Behaviour
Analyse credit reports, active loans, arrears, inquiries, debt exposure and repayment history.
Identify combinations of bureau and transactional variables that best predict defaults.
Evaluate LendGo's historical paid and defaulted loans to improve credit decisioning.
4. Credit Decisioning & Profitability
Recommend data-driven approval, referral and decline thresholds.
Optimise approval rates while maintaining sustainable collection performance.
Develop risk-based loan amount recommendations.
Analyse expected credit losses, collection performance and profitability by customer segment.
Support the development of automated credit decisioning rules.
5. Reporting & Continuous Improvement
Develop dashboards tracking approval rates, default rates, collections and loan book profitability.
Conduct monthly scorecard performance reviews.
Monitor model drift and recommend improvements.
Translate analytical findings into practical recommendations for lending operations.
Minimum Requirements
3–5+ years of experience in credit risk modelling, credit scoring or predictive analytics.
Direct experience in unsecured consumer lending, microfinance, payday lending or digital lending.
Strong Python and SQL skills.
Experience developing and validating credit risk scorecards.
Experience analysing bank transaction data and credit bureau information.
Knowledge of probability of default, credit losses and lending profitability.
Demonstrated ability to translate predictive models into operational credit decisions.
Familiarity with South African lending regulations, including the National Credit Act, is advantageous.
Key Performance Indicators
Predictive model discrimination and calibration.
Reduction in first-payment defaults.
Improvement in risk-adjusted loan profitability.
Approval rates at agreed collection-risk thresholds.
Model stability and effectiveness over time.
Measurable improvement in credit decisioning efficiency.
Our objective: Build a scalable, data-driven lending business that approves more creditworthy customers while reducing credit losses and maximising sustainable profitability.

Work arrangement
Yes