Risk Models Validation

Business Situation

In the banking industry, credit risk models have to comply to high standards in order to be compatible with banking regulations. Additionally, they have to satisfy model performance thresholds set by organizations themselves during periodical monitoring processes. To satisfy performance and regulative requirements, financial institutions have to perform validation studies. Cutting edge technologies and new algorithms opened a new era for data analytics in banking. Due to the potential of increasing model performance, the interest in AI models increased rapidly in the last decade. With this purpose our client wanted to apply a validation framework to self-learning AI Based models.

Our Approach & Solution

Our approach is to apply the validation framework to readily developed AI models. On top of validating currently performing model codes, we are challenging their code structure as well.


We implemented our validation controls via PyCharm or Spyder based on Python language.