Director, AI Security

Builds Gibson Dunn's first AI security function from scratch, defining the firm's AI security strategy with the CISO and deciding the team's shape, size, sequencing and budget. The director reports to the Chief Information Security Officer and is based in New York.

The role sets the target architecture, roadmap and control framework for AI, recruits and leads the new team, and owns security review in the AI use-case approval process, working with the Office of General Counsel on confidentiality and privilege. It reviews the design of AI platforms, RAG pipelines and agent frameworks, defines controls against prompt injection, tool abuse, data leakage and cross-matter contamination, and tackles unsanctioned AI use. A large part is identity: owning the firm's model for agent and non-human identities, keeping agents within user entitlements and ethical walls, and setting standards for MCP servers and connectors so every agent action is auditable. The director also plans AI use inside security operations, sets up AI red teaming and pre-deployment testing, owns AI incident playbooks, assesses AI and legal-tech vendors, turns regulation such as the EU AI Act into requirements, and deputises for the CISO on AI matters.

The firm wants ten or more years in information security with end-to-end ownership of a major domain, deep practical experience securing production LLM systems, strong identity foundations including OAuth, OIDC and machine identity, cloud and SaaS security architecture, and a record of building a capability from the ground up. Executive presence with partners and clients matters.