AI Validation, Senior Specialist

Vanguard•Charlotte, NC

19h•Hybrid

About The Position

The Model Risk Management (MRM) Team, part of Vanguard’s second line of defense, seeks an experienced model risk professional to provide independent oversight of AI/ML and generative AI models built and/or used across the enterprise. In this role, the Senior Specialist leads independent validation and challenge of AI models to ensure they are conceptually sound, well-governed, and fit for purpose. The role works closely with data scientists, quantitative developers, model users, and technology partners while maintaining collaborative relationships with the business. The Senior Specialist also contributes to the evolution of Vanguard’s model risk management framework by shaping methodologies, standards, and practices that support consistent, enterprise-wide application of model risk principles. The role assesses generative AI models and systems throughout their lifecycle, advises stakeholders on model risk policy, identifies emerging risks, and mentors junior team members. Vanguard’s hybrid working model applies to the majority of crew members and is designed to enable flexibility alongside in-person learning, collaboration, and connection.

Requirements

PhD in a quantitative discipline such as Computer Science, Mathematics, Statistics, Physics, or Engineering, or an equivalent combination of training and experience.
Minimum of 7 years of relevant experience in model development or model validation.
Experience developing, testing, or validating AI/ML models or systems using generative AI.
Knowledge of programming languages, including Python.
Knowledge of technology platforms used for model development and deployment, and CI/CD.
Knowledge of generative AI architectures, including foundation models, RAG systems, agentic workflows, and prompt engineering approaches, and associated risks such as hallucinations, algorithmic bias, and harmful outputs.
Knowledge of AI/ML modeling, including ML algorithms, NLP, deep learning, and evaluation methodologies.
Exceptional written communication skills, including the ability to communicate complex model risks to audiences ranging from technical practitioners to senior executives.
Ability to deal effectively with a variety of stakeholder teams.
Ability to effectively manage multiple and competing priorities.
Flexibility and excellent judgment.

Nice To Haves

Experience in the financial services sector.
Broad knowledge of risk management at financial institutions.

Responsibilities

Assess models and systems using generative AI throughout their lifecycle, including leading independent validations and assessing ongoing monitoring, change management, and remediation of model risk findings.
Provide effective challenge of model assumptions, methodologies, data, testing, implementation, monitoring, and limitations, focusing on material risk drivers.
Assess compliance with internal AI standards, emerging regulatory requirements, and governance expectations related to AI/ML and generative AI systems, including fairness, explainability, transparency, and human oversight.
Ensure timely execution of model validation and produce high-quality validation reports that meet internal standards; communicate findings clearly to technical and non-technical stakeholders.
Develop, maintain, and enhance validation procedures for a consistent, risk-based approach to independent validation and effective challenge of AI/ML and generative AI systems, keeping informed about developments in generative and agentic AI.
Contribute to the design and enhancement of model risk management policies, standards, and procedures across the model lifecycle, including inventory, development, validation, approval, and ongoing monitoring.
Advise stakeholders on Model Risk policy and other relevant standards as needed.
Identify emerging and top risks across the portfolio of AI/ML and generative AI models and surface them to MRM leadership.
Coach and mentor junior team members.
Participate in special projects and perform other duties as assigned.