About this role
Gibson Dunn needs a strategic security leader to establish and shape its inaugural dedicated artificial intelligence security function from the ground up, as there is currently no inherited team or playbook for this domain. Based in New York City, the Director of AI Security will report directly to the Chief Information Security Officer to formulate the firm's AI security strategy, target architecture, and multi-year roadmap. Working alongside executive leadership and practice groups, this individual will architect defensive controls against prompt injection, data leakage, and cross-matter contamination while governing complex agentic identity models and non-human authentication.
Beyond high-level strategy, the person in this position will conduct hands-on security architecture reviews for retrieval-augmented generation pipelines, agent frameworks, and custom internal builds. Candidates must balance deep technical knowledge regarding token scopes with the executive presence needed to engage credibly with senior partners, practice group leaders, and external clients. Anyone who excels at building security capabilities from scratch, influencing stakeholders without formal authority in a professional services setting, and translating complex regulatory frameworks into actionable controls will thrive in this role.
What you will do
- Partner with the Chief Information Security Officer to define the firm's overarching artificial intelligence security strategy, multi-year roadmap, and target architecture.
- Design the security control framework for artificial intelligence systems while overseeing the approval process for new use cases.
- Build relationships with practice group leaders and partners to translate adoption drivers into secure delivery options.
- Conduct security architecture and design reviews for platforms, retrieval-augmented generation pipelines, and agent frameworks.
- Govern non-human identities, agentic identity models, authorization patterns, and delegated access for tools and connectors.
- Develop and own the roadmap for applying automation and artificial intelligence within security operations.
- Establish red teaming programs, security evaluation criteria, and incident response playbooks tailored to artificial intelligence threats.
What you need
- Hold a bachelor's degree in Computer Science, Information Technology, Cybersecurity, or a related field, or possess equivalent professional experience.
- Demonstrate at least 10 years of information security experience that includes end-to-end ownership of a major security domain.
- Bring practical, hands-on experience securing production artificial intelligence and large language model systems against domain-specific threats.
- Possess strong foundations in identity and access management, including OAuth 2.0, OIDC, SAML, delegation patterns, and machine identity at scale.
- Show security architecture experience spanning cloud environments like Azure and AWS alongside SaaS-heavy ecosystems.
- Exhibit a track record of building security capabilities from the ground up rather than inheriting pre-existing teams.
Nice to have
- Relevant certifications such as CISSP, CCSP, CISM, or equivalents
Benefits
- Health care
- Retirement benefits
- Paid days off, including sick time and vacation time
- Parental leave
- Basic life insurance
- Flexible Spending Accounts and discretionary, performance-based bonuses
Skills and tech stack
Questions candidates ask
Where is this role located?
This position is based in New York City.
What is the compensation range for this position?
The annual compensation range for this position is $300,000 to $380,000, with the final salary depending upon qualifications and other operational considerations alongside discretionary, performance-based bonuses.
Who does this role report to?
The Director of AI Security reports directly to the Chief Information Security Officer.
Is there an existing team for this function?
This is the firm's first dedicated artificial intelligence security role, meaning there is no inherited team or playbook.