Nexus IT Group is seeking an AI Model Validation Specialist to conduct independent reviews and validation of AI and machine learning models. The role involves assessing model risks, providing governance oversight, and ensuring compliance with industry standards.
Responsibilities:
- Conduct independent reviews and validation of AI and machine learning models, evaluating:
- Model architecture and methodology
- Assumptions, limitations, and intended use
- Training, testing, and validation datasets
- Performance metrics, thresholds, and benchmarking approaches
- Review third-party model documentation, technical reports, and supporting artifacts to assess completeness, transparency, and accuracy
- Evaluate whether models are operating as intended and producing reliable, defensible outcomes
- Assess model risks across multiple dimensions, including:
- Fairness and potential bias
- Explainability and transparency
- Reliability and operational limitations
- Edge-case performance and unintended outcomes
- Evaluate adherence to industry model risk management standards and emerging AI governance expectations
- Identify control gaps, document findings, and recommend practical risk mitigation strategies
- Provide independent input into AI governance, approval, and oversight processes
- Partner with cybersecurity, technology, compliance, and risk teams to evaluate new AI use cases and model deployments
- Serve as an independent reviewer, challenging assumptions, methodologies, and conclusions to strengthen decision-making and risk management practices
- Lead periodic reviews and re-validations of AI and predictive modeling solutions
- Evaluate model monitoring programs and key performance indicators, including:
- Model drift
- Stability and consistency
- Performance deterioration
- Emerging risk trends
- Recommend remediation plans when performance, control, or risk thresholds are exceeded
- Ensure monitoring activities align with enterprise AI governance requirements
- Maintain comprehensive validation documentation and supporting evidence
- Produce validation reports that are transparent, reproducible, and audit-ready
- Support internal audits, examinations, and regulatory reviews through technical analysis and documentation
- Present validation findings to both technical and non-technical stakeholders