Staff AI Security Researcher

The role

You will lead efforts to identify and understand security risks unique to AI systems. You will investigate attacks like prompt injections and jailbreaks to build detection and response capabilities.

What you will do

• Conduct advanced research to identify vulnerabilities within AI models, agentic systems and their infrastructure
• Build threat models for emerging AI capabilities and deployment surfaces
• Innovate and design novel methodologies and automated tools to probe AI security mechanisms
• Partner with engineering and product teams to translate research findings into Agentic Detection and Response offering
• Stay informed on academic, industry and threat landscape trends on adversarial ML, agentic and AI security
• Mentor other researchers and engineers
• Publish research findings internally and externally

Required qualifications

• 6+ years of experience in offensive research, red teaming or pentesting, with focus on AppSec or Identity
• Hands-on experience with adversarial techniques against AI (LLMs, SLMs) and ML systems
• Good understanding of identity/access management (IAM) protocols and tools
• Familiarity with agentic attack taxonomies such as OWASP Top 10 for LLMs and MITRE ATLAS
• Proficiency in programming and scripting languages (Python, JavaScript and Java)
• Proficiency in Cloud environments (AWS, GCP, Azure, Kubernetes)
• Technical writing & content development skills

Nice to have

• Publication record or demonstrable body of work in security research, adversarial ML, or AI safety
• Experience in leading cross-functional technical initiatives

Benefits

• Restricted Stock Units (RSUs)
• Employee Stock Purchase Plan (ESPP)
• Competitive leave benefits
• Gender-neutral parental leave
• Employee Assistant Program
• Workout sessions and a Wellness App
• Medical and insurance benefits
• Pension
• Global home office allowance
• Mobile phone reimbursement
• Study Fund

About the company

SentinelOne is a company at the intersection of AI and security. Their AI-native platform unifies protection across endpoint, cloud, identity, data, and AI systems.