Senior AI Security Engineer

NVIDIA•Seattle, WA

2d•$184,000 - $356,500

About The Position

NVIDIA’s Agentic & AI Cybersecurity Research team is seeking a hands-on Senior Engineer to turn sophisticated problems and early prototypes into practical solutions. The role builds software to advance cybersecurity tooling for AI and collaborates with research, security, product, and partner teams to develop research code into alpha product versions and production-ready services. The engineer will identify valuable MVPs for ambiguous security and identity problems, validate prototypes with partner teams, and create reusable tools and integration patterns that help internal teams adopt secure agent approaches across enterprise tools and platforms. The role partners across Enterprise Security, Product Security, Enterprise AI, Enterprise Product, and platform teams to align architecture with real workflows and operational constraints. The position requires strong software engineering capabilities, experience in security or related areas, familiarity with agents and AI tooling, and the ability to develop ideas into production systems. NVIDIA also values human-centered problem solving, reusable engineering frameworks, open-source experience, and understanding of AI-specific security risks. The base salary range is $184,000–$287,500 for Level 4 and $224,000–$356,500 for Level 5, determined by location, experience, and pay of employees in similar positions. The role is also eligible for equity and benefits. Applications will be accepted at least until October 24, 2026. This posting is for an existing vacancy.

Requirements

Bachelor’s degree, Master’s degree, or equivalent experience in Computer Science or a related field.
8+ years in software engineering, security engineering, or related areas, including hands-on technical leadership.
Excellent software engineering skills, including building backend services, APIs, automation, or platform components in various programming environments.
Ability to quickly learn new paradigms, frameworks, and tools as needed.
Experience with multi-contributor and long-running software projects.
Familiarity with building agents, AI tooling, and architectures.
Track record of framing problem spaces, defining an MVP, prototyping quickly, and evolving ideas into production systems.
Familiarity with offensive security, adversarial thinking, red teaming, or attack-path analysis, and ability to apply that perspective to architecture decisions and security tradeoffs.
Flexibility and willingness to collaborate across organizational boundaries while balancing security with usability and adoption.

Nice To Haves

Strong human-centered thinking and ability to integrate requirements from researchers and users into code.
Excellent software engineering instincts, including building reusable frameworks, understanding and refining requirements, and creating tooling that improves engineering velocity and consistency.
Deep familiarity with open-source projects, including community involvement, workflow, and governance.
Experience with agentic or AI-adjacent systems, such as tool-using applications, workflow orchestration, evaluation loops, policy controls, memory/context services, or trust and discovery systems.
Understanding of AI-specific security risks such as injection, tool misuse, unauthorized access expansion, data exfiltration, or unsafe delegation patterns.

Responsibilities

Build leading software to accelerate and improve cybersecurity tooling for AI.
Collaborate with research and product teams to develop research code into alpha product versions.
Identify high-value MVPs for ambiguous security and identity problems, prototype solutions, validate them with partner teams, and iterate toward production-ready services.
Develop reusable skills, CLI tools, APIs, and integration patterns that help internal teams adopt secure agent approaches across enterprise tools and platforms.
Partner with Enterprise Security, Product Security, Enterprise AI, Enterprise Product, and relevant platform teams to align architecture with real workflow needs and operational constraints.