Staff Security Engineer, AI Vulnerability Research
5 Tage altAngaben zum Job
| Firma | |
| Kategorie | Informatik | Pensum | 100% |
| Einsatzort | Zürich |
Job-Inhalt
Minimum qualifications:
- Bachelor's degree or equivalent practical experience.
- 8 years of experience in security engineering.
- Experience in operational security risk assessments or vulnerability assessment.
Preferred qualifications:
- Experience in supporting security attack prevention, compliance audits and programs.
- Experience in one or more programming languages suitable for security research and prototyping (e.g., Python).
- Experience with security monitoring and analysis tools.
- Experience and interest in mentoring junior team members.
- Experience in offensive security and vulnerability research.
About the job
Our Security team works to create and maintain the safest operating environment for Google's users and developers. Security Engineers work with network equipment and actively monitor our systems for attacks and intrusions. In this role, you will also work with software engineers to proactively identify and fix security flaws and vulnerabilities.
Responsibilities
- Identify security vulnerabilities in Google's core AI and ML infrastructure through technical analysis, code review, design review, and fuzzing.
- Conduct vulnerability research into specific components and threat areas, such as model exfiltration vectors, tampering techniques, insecure input managing, or infrastructure abuse potential.
- Develop proof of concept exploits and tools to demonstrate the impact and exploitability of discovered vulnerabilities. Lead and participate in offensive security exercises (e.g., Orange Team, pen tests) against critical AI systems and infrastructure.
- Analyze complex system architectures and threat models related to AI development and deployment to identify systemic weaknesses and security gaps. Collaborate closely with AISS hardening engineers and ML infrastructure teams to design and validate effective mitigation for identified vulnerabilities.
- Document research findings, vulnerability details, exploitation techniques, and mitigation recommendations for technical audiences.