About OpenTrain
OpenTrain is the centralized, open platform where AI training and data-labeling freelancers find work, consolidate opportunities across platforms, and build a unified AI training portfolio they control. We make specialized AI training work easier to track, apply for, and grow into a durable freelance career.
About this role
OpenTrain is recruiting a GenAI Security Evaluation Engineer to create and assess code-based examples that help evaluate how effectively static analysis tools detect vulnerabilities in generative AI applications. You will work with runnable agent and retrieval-augmented generation repositories covering Sensitive Information Disclosure and Excessive Agency across tool calling, memory, and MCP. The work combines secure software engineering, vulnerability analysis, deterministic testing, and structured technical annotation.
What you will do
Build small, runnable agent and RAG repositories with code-reachable security vulnerabilities. Create vulnerable, fixed, and hard-negative variants with minimal security-relevant differences. Trace and annotate assets, data and action paths, controls, root causes, severity, and residual risk. Define authorization contexts and write deterministic tests for vulnerable, fixed, and negative behavior. Recommend security controls and contribute to calibration and peer review.
Required skills
At least 5 years of experience in application or product security, or security-focused software engineering, including secure code review. Experience with source-to-sink analysis, taint analysis, SAST, CodeQL, or Semgrep. Hands-on experience building LLM agents or RAG systems with tools such as LangChain, LlamaIndex, OpenAI or Anthropic SDKs, or MCP. Strong understanding of authorization, trust boundaries, tenants, OAuth, IAM, identity, permitted data and actions, recipients, and document-level access control. Production coding experience in Python and/or TypeScript, with sound judgment when distinguishing genuine findings from non-findings.
Schedule and compensation
This is a remote contractor engagement of up to 2 weeks. A minimum commitment of 20 hours per week is required, with at least 4 hours of overlap with Pacific Time. Compensation is up to $150 per hour, depending on evaluation and location.
Helpful background
Experience with OWASP LLM security risks, MCP, threat modeling, or security evaluation is helpful.
Why work with OpenTrain
OpenTrain gives freelancers one place to manage AI training opportunities instead of rebuilding proof of work on every platform. A stronger OpenTrain profile helps you show credible experience, discover roles that match your skills, and turn AI training and data-labeling work into a long-term portfolio.