AI Security Researcher - LLM Red Teaming & Jailbreaking Specialist- Ground Floor Opportunity

We are looking for an AI Red Teamer.
2 to 5 years of experience in AI/LLM security, adversarial testing, application security, or offensive security.
This role will focus on identifying, reproducing, and demonstrating security weaknesses in AI-powered applications, LLMs, agents, RAG systems, and AI-integrated platforms. The ideal candidate understands how modern AI systems can be abused and can think creatively about attack paths that go beyond traditional application security.
Responsibilities
Perform adversarial security testing against LLMs, AI agents, RAG systems, APIs, and AI-enabled applications.
Design and execute attacks covering the OWASP Top 10 for LLM Applications and MITRE ATLAS techniques.
Test for prompt injection, indirect prompt injection, jailbreaks, sensitive information disclosure, excessive agency, insecure tool use, improper output handling, system prompt leakage, and authorization weaknesses.
Identify attack chains involving AI agents, tools, plugins, MCP servers, APIs, external data sources, and third-party integrations.
Evaluate AI systems for cross-user or cross-tenant data exposure and privilege-boundary violations.
Assess RAG pipelines for poisoning, unauthorized retrieval, knowledge-base manipulation, and data leakage.
Evaluate AI agents for unintended autonomous actions and abuse of available tools or permissions.
Develop repeatable adversarial scenarios, proof-of-concepts, and security test cases.
Work with engineering and security teams to reproduce findings and recommend practical mitigations.
Document findings with clear technical evidence, attack paths, business impact, and remediation guidance.
Stay current with emerging AI attack techniques and evolving AI security frameworks.
Qualifications
Up to 5 years of relevant experience in AI security, red teaming, penetration testing, application security, security research, or related offensive-security roles.
Strong understanding of LLM and generative-AI architectures.
Hands-on knowledge of the OWASP Top 10 for LLM Applications.
Familiarity with MITRE ATLAS and adversarial AI techniques.
Understanding of AI agents, RAG, embeddings/vector databases, APIs, tool calling, and modern agentic architectures.
Strong knowledge of web and API security.
Ability to automate testing using Python or similar scripting languages.
Strong analytical mindset and ability to develop unconventional attack scenarios.
Ability to clearly communicate technical findings.
Good to Have
Experience researching or discovering new AI/LLM attack techniques.
Experience testing MCP-based or agentic systems.
Traditional penetration-testing or red-team background.
Participation in AI security research, bug bounty programs, CTFs, or security conferences.
Public security research, CVEs, tools, blog posts, or conference presentations.
Write to me @ rekha.cxc@careerxperts.com

Work arrangement
Hybrid