SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Eye Security is hiring a Staff AI Security Engineer to champion AI security as a discipline and independently explore how modern AI and agentic systems can make SecOps more effective and solve cybersecurity problems at scale.
You will expand Agentic SOC capabilities by working directly with SecOps analysts to identify high-value investigation, triage, and response workflows that can be improved through AI. You'll prototype agents, evaluate them against real operational cases, and measure their effect on analyst quality, speed, and workload. You will develop agentic workflows at scale—for example, researching and prototyping systems that help Eye identify relevant vulnerabilities, validate exposure and exploitability, develop detection logic, and prioritize action across customer environments.
A key part of the role is translating prototypes into customer-facing capabilities and tools. You'll actively explore with customers and partners the possibility of turning a prototype into a shippable tool, such as browser extensions, agentic skills, and other tools that solve real problems. You will also help secure Eye's AI systems by assessing agentic, LLM, and RAG systems for realistic risks such as prompt injection, unsafe tool use, data leakage, and supply-chain exposure, then turn findings into practical engineering controls.
You'll work closely with Principals in your team who have hands-on AI expertise and emphasize cooperation and mutual exchange. You'll work across the organization to get findings and mitigations into shipped systems and understand where AI-related incidents and threats are already showing up operationally. You'll have exposure to write technical research for the Eye Research blog and represent Eye Security externally through talks and technical content on AI security. You'll also track the threat landscape, staying ahead of emerging attack techniques against LLM and agentic systems and emerging guidance (OWASP, NIST, ENISA) on how to defend against them.
In your first 6–12 months, you will write and execute a 100-day plan with input from the Research & Innovation team and the CTO. You'll work toward an Agentic SOC prototype and/or AI-driven CVE hunting capability that demonstrates measurable improvement in analyst efficiency or investigation quality. You'll complete a security assessment of at least one production AI system and help implement the resulting controls, establishing a working definition of "AI security review" at Eye Security. You'll publish research or tooling when it creates meaningful value for customers or the wider security community.
REQUIREMENTS:
- Strong hands-on cybersecurity experience in one or more areas such as SOC operations, incident response, threat detection, vulnerability research, offensive security, or security engineering.
- Staff-level technical judgment: ability to independently identify valuable research questions, decide what is worth pursuing, communicate reasoning, and influence engineers and security specialists without formal authority.
- Comfort building things yourself: ability to go from a research question to a working prototype largely independently.
- Demonstrable hands-on interest in AI: experience building or experimenting with LLMs, agents, RAG, or related systems through professional work, independent research, open-source contributions, or substantial personal projects.
- Working proficiency in English. Dutch or German is a plus, not a requirement.
BONUS:
- Prior offensive security experience (pentesting, red teaming, vulnerability research) that transfers to the AI/LLM domain.
- Background in SOC operations, MDR, or incident response.
- Contributions to the AI security community: CTF write-ups, published jailbreaks or mitigations, OWASP LLM Top 10 or MITRE ATLAS involvement, conference talks.
- Experience with managed LLM platforms (e.g., Bedrock).
