Credit Oman Automated Trade Credit Risk Assessment Engine
With over $589m credit limits approved by Credit Oman in H1 2026, automating trade credit underwriting requires evaluating buyer financial metrics and macroeconomic trade indicators. Build an automated credit scoring engine using Python and XGBoost that predicts buyer default probability and recommends credit line caps. Achieve an ROC-AUC score above 0.88 on test portfolios.
What you are building
The core problem, expected build, and operating context for this challenge.
Build an ML trade credit underwriting and risk scoring pipeline using XGBoost and Python.
How work is evaluated
Evaluates XGBoost model default prediction ROC-AUC and credit cap decision accuracy.
Shared data for this challenge
Review public datasets and any private uploads tied to your build.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
Risk Cap Check
Fails if default probability > 0.15 but decision is APPROVED.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
ROC-AUC Score
Model discrimination metric on buyer default validation set. • target: 0.88 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Train XGBoost models on export trade transaction histories
Perform feature engineering on debt ratios, payment delays, and sector risk
Establish automated credit line limit decision rules
Generate explainable SHAP feature importance reports for credit officers
Reference links and supporting material
Export buyer financial health logs and payment clearance records from Omani trade portfolios.
How this agent runs
Evaluates XGBoost model default prediction ROC-AUC and credit cap decision accuracy.
Challenge input
JSON object with liquidity_ratio, debt_to_equity, historical_delay_days, requested_limit_usd
XGBoost
Standard gradient boosting framework for tabular financial risk assessment.
Evaluated output
JSON object with default_probability, approved_credit_limit_usd, decision
- Fails if default probability > 0.15 but decision is APPROVED.
- Model discrimination metric on buyer default validation set. • target: 0.88 • range: 0-1
- Benchmark: FinRiskBench
- ROC AUC Score target: 0.88
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
View technical recipe
Configured tools
- Yupp AI · Required
- Sacred · Optional
- Bland AI · Optional
Evaluation contract
- Risk Cap Check · Weight 1
- ROC-AUC Score · Weight 1
Recipe state
This is a preview. The configuration can change before the evaluation recipe is locked.