Program Overview
Financial institutions need faster, more precise decisions as credit risk and fraud techniques grow increasingly complex. Traditional rule-based systems struggle to keep up, making AI/ML essential for accurate scoring, real-time detection, and reliable portfolio insights. This program equips professionals to evaluate and apply ML models across lending and fraud workflows, interpret outputs responsibly, and align decisions with regulatory expectations. Through practical use cases and simulation-based exercises, participants learn to use machine learning to enhance accuracy, reduce losses, and support stronger governance.
Features
- Interpret and apply ML scoring and fraud models to lending and risk decisions
- Evaluate feature inputs, performance metrics, thresholds, and model reliability
- Strengthen governance, fairness, and explainability for regulatory trust
- Use model outputs to refine approval policies, fraud controls, and portfolio strategy
Target audiences
- Risk, Credit & Lending Teams
- Fraud, Financial Crime & Operations
- Data Science & Analytics Professionals
- Product, Strategy & Digital Banking Teams
Curriculum
- 4 Sections
- 19 Lessons
- 1 Day
Expand all sectionsCollapse all sections
- AI/ML Foundations for Regulated Financial Decisioning5
- 1.1ML vs rule-based decisioning: AI enhances risk accuracy and efficiency
- 1.2Core ML model categories: Classification, anomaly detection, supervised vs unsupervised learning
- 1.3Feature engineering, model training, testing, validation, and drift
- 1.4Key evaluation metrics: accuracy, recall, ROC-AUC, precision, false positives/negatives
- 1.5Explainability requirements (XAI), fairness, bias mitigation, and alignment with supervisory expectations
- Practical Application for Credit Scoring & Underwriting5
- 2.1Traditional vs ML-based scoring frameworks: FICO-style vs behavioral and alternate-data models
- 2.2Feature inputs: Bureau data, financial statements, behavioral patterns, device signature signals, employment and segmentation attributes
- 2.3Probability of Default (PD), Loss Given Default (LGD), and risk segmentation modeling
- 2.4Loan pricing and decision policy optimization using model outputs
- 2.5Portfolio recalibration, population stability analysis, and ongoing model risk management
- Real-World Fraud Detection and Behavioral Risk Modelling5
- 3.1Real-time payment fraud detection using anomaly detection and NLP features
- 3.2Synthetic identities, mule accounts, and high-risk behavioral markers
- 3.3Card transaction risk scoring, merchant risk analysis, and cross-channel fraud prevention
- 3.4AI-enabled AML transaction monitoring: pattern analysis and hidden signal discovery
- 3.5Supervisory and audit expectations: control testing, governance, and failure case remediation
- Simulation Labs & Decisioning Experiments4
- 4.1Model Interpretation Lab: Evaluate model outputs and adjust thresholds for accuracy–risk balance
- 4.2Credit Portfolio Simulation: Re-score a portfolio to assess changes in approval, pricing, and risk
- 4.3Fraud Decision Exercise: Prioritize alerts using severity and likelihood scoring
- 4.4Bias & Explainability Review: Detect fairness gaps and prepare a regulator-ready explanation
Instructor
