Perplexity is seeking a highly skilled, experienced and hands-on Offensive Security Engineer to join our dynamic security team, taking an adversarial approach to hardening Perplexity's infrastructure, applications, and AI systems. You'll plan and execute red team operations, penetration tests, and attack simulations across our cloud infrastructure, web and mobile applications, AI/ML pipeline, and corporate environment—finding real vulnerabilities before adversaries do and working directly with engineering teams to drive remediation.
Responsibilities
Plan and execute red team and purple team engagements simulating advanced threat actors across cloud infrastructure (AWS, Kubernetes), endpoints, and application surfaces
Conduct continuous penetration testing of web applications, APIs, mobile clients, browser extensions, cloud infrastructure, and internal services
Assess AI/ML-specific attack surfaces including prompt injection, model exfiltration, agent abuse, tool-use exploitation, and MCP security boundaries
Develop and maintain custom offensive tooling, exploits, and automation to improve the efficiency and coverage of security testing
Perform open-scope adversary simulations that test detection and response capabilities end to end, collaborating closely with the defensive security team
Drive threat modeling sessions with engineering teams to identify and prioritize attack vectors in new features and architectures
Deliver clear, actionable findings to both technical and executive audiences; partner with engineering to validate remediations
Contribute to the security of CI/CD pipelines, supply chain integrity, and secrets management through offensive assessment
Stay current on emerging attack techniques, vulnerability research, and adversary tradecraft; bring external perspective into Perplexity's security strategy
Qualifications
5+ years of hands-on experience in offensive security, red teaming, or penetration testing
Deep technical expertise in at least two of: cloud security (AWS/GCP/Azure), web/API application security, Kubernetes and container security, macOS/Linux endpoint security, network penetration testing, or CI/CD pipeline security
Track record of discovering impactful vulnerabilities or developing novel attack techniques in production environments
Strong programming and scripting skills in Python, Go, or similar languages; comfortable writing custom tooling and exploits
Experience with industry-standard offensive tools (Burp Suite, Cobalt Strike / Sliver / Mythic, Metasploit, BloodHound, nuclei, etc.) and ability to operate beyond them
Excellent written and verbal communication; able to translate complex technical findings into clear risk narratives
Experience assessing AI/ML systems, LLM applications, or agentic workflows for security vulnerabilities
Bonus: Published security research, conference talks (DEF CON, Black Hat, BSides), CVE credits, or meaningful bug bounty contributions