Advanced Risk Modeling for Smarter Decisions.
Quickly gain an intuitive understanding of the data.

We develop advanced credit risk solutions and methodologies for financial institutions within the Romanian banking system.
The framework is applicable to the construction of a wide class of probabilistic models and more.

Model Studio automatically generates candidate linear and non-linear predictive models, including classical linear models (OLS) for regression, logistic regression / SDCA for classification, and non-linear models (e.g. FastTree). The models are compared in a leaderboard, and clear explanations are provided: coefficients (top 50 by absolute value) for linear models and feature importance using Permutation Feature Importance (PFI) for non-linear models, while a logistic scorecard model is presented in detail in the Scorecard section.

We develop credit risk models covering all key areas, including PD, LGD, and EAD modeling, IFRS 9 ECL computation, staging and SICR assessment, macroeconomic linkage and scenario analysis, as well as Basel-compliant rating and calibration frameworks. Our models are designed to meet both regulatory and accounting requirements, with a strong focus on robustness, explainability, and audit readiness.