[UDCF6VCV] Senior AI Security Engineer – LLM Vulnerability
Summary
Short-term remote freelance/contractor role via ZettaMine Labs (working with Turing): identify, document and test security vulnerabilities in GenAI/LLM applications — building vulnerable and patched app versions, assessing RAG and agent systems with tools like CodeQL and Semgrep, and writing exploit/reproduction test cases.
Hello,
Greetings from ZettaMine!!!
Freelance / Contractor Opportunity: We are hiring Senior AI Security Engineers – LLM Vulnerability for a short-term remote freelance/contractor assignment with Turing.
Locations: Bangladesh, Brazil, Colombia, Egypt, Ghana, India, Indonesia, Nigeria, Turkey, USA.
Role: Senior AI Security Engineer – LLM Vulnerability
Experience: Minimum 3+ Years
Mode: Remote
Start Date: Immediate / ASAP
Mandatory Eligibility Criteria
Role Overview
We are looking for experienced GenAI Security Engineers to create, evaluate and test security vulnerabilities in Generative AI applications. Candidates will work on identifying security weaknesses, developing vulnerable and patched application versions, documenting security findings, and creating test cases to validate vulnerabilities and their fixes.
This opportunity is ideal for professionals with experience in Application Security, Product Security, Penetration Testing, Secure Code Review and AI/LLM Security.
What does day-to-day look like:
- Source or develop GenAI applications, including AI agents and Retrieval-Augmented Generation (RAG) systems.
- Identify and create vulnerable application versions along with their corresponding patched versions.
- Perform security assessments and document findings with exact code references, vulnerability details and recommendations.
- Analyze data flows and action flows across files, callbacks, registries and background jobs to identify security gaps.
- Evaluate authorization mechanisms, authentication, tenant isolation, trust boundaries and access control.
- Identify vulnerabilities related to sensitive information disclosure, prompt injection, excessive agency and tool abuse.
- Develop exploit and reproduction test cases to demonstrate vulnerabilities and validate security patches.
- Map identified vulnerabilities to relevant CWE categories and security standards.
- Evaluate LLM application security using tools such as CodeQL and Semgrep.
- Work with LLM agents, tool calling, MCP, RAG, memory systems and tracing frameworks.
- Demonstrate security vulnerabilities using live AI models wherever possible.
Requirements:
- Experience: Minimum 3+ years in Application Security, Product Security, Secure Code Review, Penetration Testing or Vulnerability Research.
- Security Knowledge: Strong understanding of authentication, authorization, least privilege, tenant isolation, trust boundaries and approval workflows.
- LLM Security: Knowledge of OWASP Top 10 for LLM Applications, especially Sensitive Information Disclosure (LLM02) and Excessive Agency (LLM03).
- Vulnerability Assessment: Experience identifying prompt injection, tool abuse, access control issues and sensitive data exposure.
- Code Analysis: Strong ability to trace data flows and identify vulnerabilities across multiple files, callbacks and background processes.
- Security Tools: Hands-on experience with CodeQL, Semgrep, SARIF and pattern-based security scanning.
- AI Technologies: Experience with LLM tool calling, AI agents, MCP, RAG, memory and tracing.
- Frameworks: Familiarity with LangChain, LlamaIndex, OpenAI SDK, Anthropic SDK and OpenTelemetry.
- Programming: Strong production coding experience and engineering fundamentals.
- Development Tools: Experience with Docker, Git, security patches and commit history analysis.
- Communication: Strong technical English communication and documentation skills.
- Testing: Ability to write exploit and reproduction tests to validate vulnerabilities and confirm security fixes.
Good to Have:
- Experience in GenAI application security and AI red teaming.
- Hands-on experience securing AI agents and enterprise RAG applications.
- Familiarity with OWASP security guidelines and CWE vulnerability classifications.
- Experience reviewing upstream security patches and identifying logic-level vulnerabilities.
- Knowledge of secure AI architecture and authorization design.
Availability Requirements:
- Full-time or Part-time: 40 hours or 20 hours per week.
- Mandatory: Minimum 4 hours of PST timezone overlap.
- Candidates should be available to start immediately.
Interested Candidates:
If you meet the eligibility criteria and are interested in this remote freelancing opportunity, please share your updated CV and LinkedIn profile at md.afreen@zettamine.com.
Thanks & Regards,
Saranya M (Afreen)
ZettaMine Labs Pvt. Ltd.
Skills
Sign in to read and join the discussion.