Founding Security Researcher

Join an early-stage cybersecurity company as its first AI security researcher. You will lead threat research into AI models and agents, publish findings that inform the security community, and help shape product protections for enterprise environments.

• Investigate agent and model harnesses for prompt injection, tool and MCP abuse, and credential or data exfiltration.
• Benchmark open-source and closed-source models against prompt injection and other attack techniques.
• Publish research and findings that build understanding of AI security risks.
• Translate research into product detections and enforcement policies, and investigate agent incidents in customer environments.
• Track emerging agent frameworks, rogue agent behavior, and open-source model risks, and brief security leaders on relevant findings.

• Two to ten years of security or threat research experience focused on AI models, with work in offensive security, red teaming, detection engineering, or AI security research.
• Hands-on experience attacking AI models or agent harnesses, including prompt injection, jailbreaks, or tool and MCP abuse.
• Background on a cybersecurity team, enterprise security team, or AI lab security team; AI or machine learning research without security experience is not a fit.
• Experience beyond vulnerability management, penetration testing, or application security alone.
• Research publications, CVEs, conference talks, or CTF results are valued.

Base salary: $175,000 to $225,000 annually. Visa sponsorship is not available. Equity and benefits details were not provided.

On-site, five days per week in San Francisco, California, United States.