Model Risk Analyst job at CRDB

Job Purpose

The Model Risk Analyst assesses and challenges the fitness, limitations and risks of AI and machine learning models throughout their lifecycle. The role performs validation assignments separate from model development, reviews monitoring evidence and tracks findings under the Manager, AI Governance and Model Risk. It supports informed governance decisions without assuming model ownership, deployment authority or formal independent assurance where the required separation has not been established.

Principle Responsibilities

• Maintain a risk-based validation schedule with the Manager, covering new models, material changes, periodic reviews and third-party systems.
• Agree intended use, scope, acceptance criteria and required evidence with relevant stakeholders before validation begins.
• Declare conflicts of interest and avoid validating models personally developed or materially modified; escalate independence constraints.
• Examine data provenance, representativeness, labels, feature timing and development/testing separation; identify leakage and unsupported assumptions.
• Reproduce performance calculations and compare results with suitable benchmarks, business thresholds and relevant customer or channel segments.
• Test model stability, calibration, explainability, fairness and sensitivity using proportionate methods and documented limitations.
• Evaluate generative AI outputs for factual accuracy, grounding, harmful bias, robustness and misuse, including reachable prompt-injection scenarios.
• Review evidence for human oversight, monitoring, auditability, privacy and security controls with the accountable specialist functions.
• Assess vendor model evidence critically; identify limitations caused by restricted model access, missing data or unverified supplier claims.
• Produce evidence-linked validation reports with findings, severity, recommendations and a clear distinction between tested and untested matters.
• Track remediation and independently retest assigned findings; escalate unresolved material risks and overdue actions to the Manager.
• Review ongoing model performance, drift, incidents and overrides; recommend revalidation, restrictions, retraining or retirement where warranted.
• Supply accurate model-risk records and committee inputs; report validation coverage, overdue reviews, finding age and repeat defects.
• Support the formally designated independent review function and governance decisions; do not approve own findings for deployment or replace formal approval.
• Assess compliance with the Bank’s AI Governance Framework, Model Risk Management Standards, Responsible AI Principles, Data Governance Requirements, and applicable regulatory obligations, and report any material gaps or control weaknesses.
• Review model monitoring frameworks, performance thresholds, alert mechanisms, and escalation procedures to ensure ongoing model effectiveness and timely identification of emerging risks.
• Support model inventory management by maintaining accurate records of model classifications, risk ratings, validation status, findings, approvals, and lifecycle reviews.
• Collaborate with Data Scientists, Model Owners, Risk Management, Compliance, Internal Audit, and Technology teams while maintaining appropriate independence in validation activities

Qualifications Required

• Bachelor's degree in Statistics, Mathematics, Data Science, Computer Science, Econometrics, Risk Management, Artificial Intelligence, Machine Learning, Quantitative Finance, Analytics, or a related discipline.
• A Master's degree in Statistics, Mathematics, Data Science, Artificial Intelligence, Machine Learning, Econometrics, Quantitative Finance, Risk Management, Computer Science, or a related field will be an added advantage.
• Professional certifications in Risk Management, Model Risk Management, Data Science, Artificial Intelligence, Machine Learning, Statistics, Financial Risk Management (FRM), Professional Risk Management (PRM), or related disciplines will be an added advantage.
• Postgraduate study or relevant model-risk, statistical or responsible AI certification is an advantage

Experience

• Minimum of 3 years of experience in relevant model validation, quantitative analysis or model-risk experience.
• Strong evidence assessment, report writing, constructive challenge and banking-risk awareness.
• Experience validating machine learning models, predictive analytics models, statistical models, scorecards, forecasting models, or AI-enabled decision systems in regulated or enterprise environments.
• Knowledge of model risk management frameworks, validation methodologies, model governance practices, and model lifecycle management.
• Understanding of responsible AI principles, fairness testing, explainability techniques, bias assessment, model monitoring, and AI governance requirements.
• Working proficiency in Python or R and SQL; ability to reproduce model results.
• Knowledge of statistical testing, calibration, stability, explainability, bias and AI evaluation.

• Maintain a risk-based validation schedule with the Manager, covering new models, material changes, periodic reviews and third-party systems.
• Agree intended use, scope, acceptance criteria and required evidence with relevant stakeholders before validation begins.
• Declare conflicts of interest and avoid validating models personally developed or materially modified; escalate independence constraints.
• Examine data provenance, representativeness, labels, feature timing and development/testing separation; identify leakage and unsupported assumptions.
• Reproduce performance calculations and compare results with suitable benchmarks, business thresholds and relevant customer or channel segments.
• Test model stability, calibration, explainability, fairness and sensitivity using proportionate methods and documented limitations.
• Evaluate generative AI outputs for factual accuracy, grounding, harmful bias, robustness and misuse, including reachable prompt-injection scenarios.
• Review evidence for human oversight, monitoring, auditability, privacy and security controls with the accountable specialist functions.
• Assess vendor model evidence critically; identify limitations caused by restricted model access, missing data or unverified supplier claims.
• Produce evidence-linked validation reports with findings, severity, recommendations and a clear distinction between tested and untested matters.
• Track remediation and independently retest assigned findings; escalate unresolved material risks and overdue actions to the Manager.
• Review ongoing model performance, drift, incidents and overrides; recommend revalidation, restrictions, retraining or retirement where warranted.
• Supply accurate model-risk records and committee inputs; report validation coverage, overdue reviews, finding age and repeat defects.
• Support the formally designated independent review function and governance decisions; do not approve own findings for deployment or replace formal approval.
• Assess compliance with the Bank’s AI Governance Framework, Model Risk Management Standards, Responsible AI Principles, Data Governance Requirements, and applicable regulatory obligations, and report any material gaps or control weaknesses.
• Review model monitoring frameworks, performance thresholds, alert mechanisms, and escalation procedures to ensure ongoing model effectiveness and timely identification of emerging risks.
• Support model inventory management by maintaining accurate records of model classifications, risk ratings, validation status, findings, approvals, and lifecycle reviews.
• Collaborate with Data Scientists, Model Owners, Risk Management, Compliance, Internal Audit, and Technology teams while maintaining appropriate independence in validation activities

• Strong evidence assessment
• report writing
• constructive challenge
• banking-risk awareness
• Experience validating machine learning models, predictive analytics models, statistical models, scorecards, forecasting models, or AI-enabled decision systems in regulated or enterprise environments.
• Knowledge of model risk management frameworks, validation methodologies, model governance practices, and model lifecycle management.
• Understanding of responsible AI principles, fairness testing, explainability techniques, bias assessment, model monitoring, and AI governance requirements.
• Working proficiency in Python or R and SQL; ability to reproduce model results.
• Knowledge of statistical testing, calibration, stability, explainability, bias and AI evaluation.

• Bachelor's degree in Statistics, Mathematics, Data Science, Computer Science, Econometrics, Risk Management, Artificial Intelligence, Machine Learning, Quantitative Finance, Analytics, or a related discipline.
• A Master's degree in Statistics, Mathematics, Data Science, Artificial Intelligence, Machine Learning, Econometrics, Quantitative Finance, Risk Management, Computer Science, or a related field will be an added advantage.
• Professional certifications in Risk Management, Model Risk Management, Data Science, Artificial Intelligence, Machine Learning, Statistics, Financial Risk Management (FRM), Professional Risk Management (PRM), or related disciplines will be an added advantage.
• Postgraduate study or relevant model-risk, statistical or responsible AI certification is an advantage