Cybersecurity SME AI Red Team
Job Description
JSM Consulting India is hiring a Cybersecurity SME AI Red Team to evaluate AI models and coding agents for cybersecurity risks and unsafe capabilities. In this remote role, you will run controlled tests in sandboxed environments, compare outputs to safety specifications, and produce actionable findings that connect technical issues to recognized security and AI risk frameworks.
What you will do
- Red-team AI coding agents in sandboxed environments using iterative testing techniques, evaluating whether models can generate offensive tools such as exploits, malware, and attack tooling.
- Assess model outputs against defined safety specifications, identify safety violations, determine severity, and flag when additional testing is required.
- Define and refine criteria for unsafe capability uplift, including handling ambiguous cases and gaps in existing safety specifications.
- Test model capabilities related to vulnerability discovery, exploit development, and malware analysis, verifying whether generated tools function as intended.
- Investigate agent-specific vulnerabilities as testing expands into agent and harness environments, including prompt injection, tool misuse, and privilege escalation.
- Create realistic attack scenarios and evaluation datasets by reviewing expert-written prompts and model outputs for technical accuracy.
- Document security findings with severity, reproduction steps, and recommended mitigations, then communicate results to engineering and safety teams.
- Map findings to established frameworks, including MITRE ATLAS, OWASP Top 10 for LLM Applications, and NIST AI Risk Management Framework (AI RMF).
- Train and mentor red teamers and annotators on cybersecurity concepts while staying current on emerging AI attack techniques.
What you bring
- Recognized certifications or proof of skill, such as OSCP, OSEP, GPEN, or evidence like published CVEs, CTF results, or a bug-bounty record.
- Prior experience in AI/ML security, adversarial ML, or model evaluation.
- Experience in government, defense, intelligence, or critical-infrastructure security.
- Background in malware reverse engineering, cloud security (AWS, GCP, Azure), or secure software development.
- Conference talks, research papers, or open-source security tooling experience.
- 2+ years in offensive security, including penetration testing, red teaming, vulnerability research, or exploit development.
- Hands-on offensive tradecraft, including exploit development, malware, post-exploitation, and web and cloud attacks, plus scripting ability (Python preferred) to validate whether model-generated tools actually work.
- Working understanding of how LLMs and agents work, including common attack classes against them.
- Strong written communication to explain technical risk to researchers and executives.
Tools and frameworks
- OSCP, OSEP, GPEN, Python, AWS, GCP, Azure
- MITRE ATLAS, OWASP Top 10 for LLM Applications, NIST AI Risk Management Framework (AI RMF)
Role details
- Location: Remote
- Experience: 2+ years
- Pay: USD 65.00 to 70.00 per hour
- Expected hours: 40.0 per week
Similar Jobs
A