At Birbank, I lead independent validation of credit risk, fraud detection, decisioning, and machine-learning models. I designed the bank’s model validation framework and standards, incorporating explainability, bias, robustness, and lifecycle monitoring.
I documented 32 material model-validation findings and assess model reliability through benchmarking, backtesting, stability testing, and failure analysis. I present validation results and model-risk assessments to senior stakeholders and the Model Risk Committee.
At Kapital Bank, I developed credit-risk scorecards used in retail lending decisions covering 64% of the portfolio, along with fraud detection models that contributed to a 12% reduction in fraud losses. I’m now expanding into applied NLP, LLM systems, RAG, and automated AI evaluation workflows.
