GenAI Security Evaluation Engineer

Create and evaluate security tasks for GenAI applications.

About Turing

Based in San Francisco, California, Turing is the world’s leading research accelerator for frontier AI labs and a trusted partner for global enterprises deploying advanced AI systems. Turing supports customers in two ways: first, by accelerating frontier research with high-quality data, advanced training pipelines, plus top AI researchers who specialize in coding, reasoning, STEM, multilinguality, multimodality, and agents; and second, by applying that expertise to help enterprises transform AI from proof of concept into proprietary intelligence with systems that perform reliably, deliver measurable impact, and drive lasting results on the P&L.

Offer Details

• Required commitment: At least 4 hours per day, minimum 20 hours per week, with 4 hours of overlap with Pacific Time.
• Engagement type: Contractor
• Engagement length: Up to 2 weeks

About the Role

We’re looking for experienced GenAI Security Engineers to create and evaluate security tasks for GenAI applications.

As a Security Task Engineer, you’ll source or build GenAI applications, create vulnerable and patched versions, document security findings, and write tests for each vulnerability mechanism. Where possible, you’ll also demonstrate the relevant behavior using a live model.

What You’ll Do

• Source or build GenAI applications, including agent and RAG systems.

• Create vulnerable versions and patched counterparts.

• Document findings with exact code line spans, traps, control status, and recommendations.

• Trace data and action flows across files, callbacks, registries, and background jobs; explain when and why a control misses a route.

• Define authorization contexts, including actors, trust boundaries, tenants, permitted data and actions, purposes, recipients, and document-level access.

• Write per-mechanism exploit and reproduction tests that demonstrate the weakness in the vulnerable version and confirm the patch prevents it.

• Demonstrate the relevant model path using a live model when possible.

What We’re Looking For

• 3+ years in application or product security, secure code review, penetration testing, or vulnerability research.

• Strong data-flow reasoning and the ability to trace a source to a sink across files, callbacks, registries, and background jobs.

• Strong authorization knowledge, including authentication, authorization, least privilege, tenant isolation, trust boundaries, and approval flows.

• Knowledge of the OWASP Top 10 for LLM Applications, especially LLM02 Sensitive Information Disclosure and LLM03 Excessive Agency. Familiarity with prompt injection and tool abuse.

• Ability to map findings to relevant CWEs, such as CWE-200/201 for exposure, CWE-532 for sensitive data in logs, CWE-284/285/862/863 for access control, and CWE-250/269 for excessive privilege.

• Experience writing exploit and reproduction tests that distinguish vulnerable behavior from patched behavior.

• Experience using CodeQL or Semgrep, reading SARIF, and understanding why pattern-based scanners can miss logic and authorization flaws.

• Hands-on experience with LLM tool calling, agents, MCP, RAG, memory, and tracing, using tools such as LangChain, LlamaIndex, OpenAI or Anthropic SDKs, and OpenTelemetry.

• Production coding experience and engineering fundamentals, including Docker, Git, reading upstream security patches and commit history, and clear technical English.
  
  

  Nice to Have

  • SWE, OSCP, GWAPT, CSSLP, or similar credentials.

  • CVE credits, bug bounty experience, or published security research.

  • Experience with AI red teaming or LLM evaluation.
    
    

Skills: Cyber Security, LLM, Code Reviews, Code Analysis