Senior Offensive Security Engineer: Autonomous Red Teaming & AI Agents (Equity)

NVIDIA Corporation; Austin (TX); On-site; USD 224,000 - 431,000; Full time

Benefits offered by this job

Equity

Benefits package

Job summary

NVIDIA Corporation is seeking a Senior Offensive Security Engineer within its Product Security organization in Austin, Texas. You will advance AI-driven offensive security by building autonomous red team agents, extending multi-agent harnesses, and encoding operator tradecraft into prompts, tools, and playbooks.

You will mentor engineers, evaluate frontier models, and shape safety with sandboxing and scope controls.

Qualifications

• Bachelor's degree or equivalent experience.
• 12+ years in security engineering with 4+ years in offensive security.
• Deep, hands-on exploitation skill in at least one domain.
• Strong Python software engineering ability; ship production code.
• Experience building with LLMs: agent frameworks, multi-agent orchestration.
• Judgment on autonomous offensive tooling: scoping and blast radius.
• Responsible approach to offensive testing; safe and fast fixes.

Responsibilities

• Craft and build new red team agents: autonomous and semi-autonomous LLM agents.
• Extend our agent harnesses across runtimes and model providers.
• Encode operator tradecraft into prompts, tools, and playbooks.
• Build verification layers to prove findings before human review.
• Develop safety controls: sandboxing, allowlists, credential isolation.
• Benchmark frontier models and mentor engineers on the team.

Skills

Python

Penetration testing

Red teaming

Exploit development

Vulnerability research

LLM agents

Education

Bachelor's degree or equivalent experience

Tools

Git

Kubernetes

OpenShift

Job description

NVIDIA Corporation is seeking a Senior Offensive Security Engineer within its Product Security organization in Austin, Texas. You will advance AI-driven offensive security by building autonomous red team agents, extending multi-agent harnesses, and encoding operator tradecraft into prompts, tools, and playbooks.

You will mentor engineers, evaluate frontier models, and shape safety with sandboxing and scope controls.