- Amsterdam, North Holland, Netherlands
- On-site
- Full-time · Mid
- Permanent contract
- 3+ years
Build and improve AI agents that help uncover security vulnerabilities during authorised penetration tests, own projects end-to-end, and develop reliable, maintainable software across LLM applications, agent harnesses, APIs, prompts, and data models.
Responsibilities
- Build and improve AI agents that help uncover security vulnerabilities during authorised penetration tests.
- Own projects end-to-end, from understanding a problem and exploring solutions to developing, releasing, and improving them in production.
- Evaluate agent performance, using real-world pentest results and feedback to understand what works and where improvements can be made.
- Develop reliable, maintainable, and debuggable software across LLM applications, agent harnesses, APIs, prompts, and data models.
- Investigate and resolve production issues, helping our AI systems become more reliable and effective over time.
Requirements
- Nice to have: AI agent harnesses
- Nice to have: ethical hacking
- Nice to have: open-source AI projects
- Nice to have: research publications
