Data Scientist Model Validation

Conduct independent validation of credit risk models used across the customer lifecycle including behavioural risk models and AI/ML-based credit decisioning models

Assess conceptual soundness, model methodology, assumptions, variable selection, feature engineering techniques, and model design.

Help establish and maintain a model inventory covering all models in production, development, and retirement stages, including risk tier and metadata management.

Define and implement model materiality classification criteria (High / Medium /Low) to prioritize validation activities and resource allocation.

Support development of MRM policies, procedures, and standards aligned with SR 26-2 (the current interagency guidance issued April 2026, superseding SR 11-7) from inception, applying its principles-based, risk-proportionate approach.

Execute validation using champion-challenger framework, benchmarking, stress testing, and sensitivity analysis.

Validate model implementation by verifying that production code and system outputs match approved model specifications.

Assess data quality, data lineage, and feature appropriateness, including reviewof credit bureau data usage.

Define ongoing monitoring standards including PSI, CSI, Gini/KS tracking thresholds, and escalation triggers for model performance degradation.  Review periodic model performance monitoring reports 

Evaluate fairness, bias, and model transparency considerations particularly for AI/ML models used in credit decisioning under ECOA/Reg B.

Author end-to-end validation reports covering scope, data assessment, methodology review, findings, risk ratings, compensating controls, and formal recommendations (Approve / Conditionally Approve / Reject).

Assign and justify model risk ratings based on model materiality, complexity, intended use, and potential business impact.

Review model development documentation, implementation documents, monitoring reports, and change management records.

Ensure all MRM activities comply with SR 26-2, OCC Bulletin 2026-13,

ECOA/Reg B, and FCRA regulatory requirements. Apply SR 26-2 &risk-based,proportionate approach to validation prioritization and governance design.

Partner with model development, risk, compliance, and technology teams to discuss findings and agree on remediation plans and model change management processes.

Present validation results to senior management, model governance committees,and client stakeholders.

Support external audits, regulatory examinations, and client due diligence activities.