Engineering Manager, Agent & Product Security
Remote • New York • FullTime
Posted 2mo ago
Remote Work Policy
Fully remote
Employment Type
FullTime
Categories
AI Agent Engineer
About the job
Cursor is building the future of coding automation, starting with a tool for professional programmers. Our mission is to make coding more efficient through inventive research, design, and engineering. We are a flat organization with a small, talented team that values truth-seeking, passion, and creativity. We thrive on spirited debate, bold ideas, and shipping code.
In this role, you will lead the Agent & Product Security team, responsible for making agentic software development trustworthy. This team defines and builds the security primitives for autonomous agents, including identity, delegated authority, policy enforcement, secure tool execution, and prompt-injection containment. These decisions will significantly impact Cursor's product direction and the industry's approach to AI security. Security is deeply integrated with our product strategy, directly influencing the pace at which we can introduce new autonomous agent capabilities.
Responsibilities
- Define security primitives for agent autonomy and lead their development.
- Engineer isolation mechanisms for secure operation within customer codebases.
- Build the identity and authority infrastructure for autonomous agents.
- Develop customer-facing security controls that are understandable and robust.
- Hire, coach, and grow a high-performing security engineering team.
- Write and review code to stay technically close to the work and ensure fast, grounded execution.
Requirements
- Experience leading engineering teams shipping production systems with high security, reliability, or platform stakes.
- Strong product and systems judgment, with the ability to reason from user workflows to architecture and threat models.
- Understanding of modern application security, authorization, tenancy, identity, secrets, data isolation, and secure-by-default platform design.
- Good instincts about AI agents, their tool usage, authority, prompt injection, and mitigation strategies.
- Comfort operating in ambiguity and setting strategy with thin precedent.
- Ability to partner effectively across product, infrastructure, ML, and customer-facing teams.