AI Red Teamer

Simbian AI is building the world's first autonomous AI security agent - one that thinks, hunts, and responds like a seasoned security analyst, at machine speed. We're backed by top-tier investors and trusted by security teams who are drowning in alerts. We're now building an AI pentesting agent that autonomously probes, exploits, and reports vulnerabilities across modern infrastructure. You'll be the adversarial brain behind it. The Role We're looking for an AI Red Teamer who understands both sides of the fence: how AI systems can be attacked, and how to build agents that do the attacking. You'll red team our own AI systems, design adversarial test scenarios, and directly shape how our pentesting agent thinks and acts. This is not a traditional red team role. You'll work at the intersection of offensive security and AI - a space most people are still figuring out

- Design and execute red team campaigns against AI/ML systems: prompt injection, jailbreaks, model extraction, adversarial inputs, training data poisoning
- Build and maintain adversarial test suites for our AI pentesting agent
- Simulate real-world attacker TTPs to stress-test agent decision-making and guardrails
- Research emerging attack vectors against LLM-based systems and agentic pipelines
- Collaborate with engineering to harden the AI agent against evasion and manipulation
- Document findings with clear severity ratings, reproducible POCs, and remediation guidance
- Contribute to Simbian's red team playbooks and threat m

Requirements

- 3-6 years in offensive security, red teaming, or penetration testing
- Hands-on experience with at least one of: LLM security, adversarial ML, or AI system testing
- Strong fundamentals in web app, API, and network pentesting (OWASP, MITRE ATT&CK)
- Scripting proficiency: Python required; familiarity with LangChain, LlamaIndex, or similar AI frameworks is a strong plus
- Working knowledge of prompt injection, jailbreaking techniques, and OWASP Top 10 for LLMs - CEH, OSCP, or equivalent - preferred, not required
- Strong written communication: your POCs need to be readable by both engineers and executives

Nice to Have
- Experience red teaming agentic AI systems or multi-agent pipelines
- Background in threat intelligence or APT simulation
- Contributions to bug bounty pro

Benefits

-Work on a problem most security teams haven't even defined yet
- Direct impact on product - your findings shape what we build
- Remote-first, async-friendly culture
- Competitive comp + equity at an early-stage AI security company - Work alongside people who've shipped security products at sc