About the role
Gramian Consulting Group is recruiting a Senior Application Security Engineer for a remote GenAI security evaluation and AI training project. The position focuses on testing how well static analysis products find application vulnerabilities in Generative AI software, especially LLM agents and Retrieval-Augmented Generation (RAG) systems. The engineer will produce compact, executable code repositories that demonstrate Sensitive Information Disclosure (SID) and Excessive Agency (EA), together with corrected examples and closely related cases that should not be flagged.
What you will do
- Produce executable agent and RAG repositories in which Sensitive Information Disclosure and Excessive Agency flaws can be reached through the code.
- Construct credible security cases using agent memory, tool calling, Model Context Protocol (MCP), and other elements of LLM applications.
- Prepare vulnerable, remediated, and hard-negative implementations whose security-related distinctions are kept minimal.
- Test static application security testing (SAST) products for accurate detection of real flaws and avoidance of false alerts.
- Trace sources to sinks, following pertinent information flows, action routes, and boundaries of trust.
- Record findings with the impacted assets, underlying causes, safeguards, severity, and remaining risk.
- Specify authorization conditions for actors, tenants, identities, allowed information, actions, purposes, and recipients.
- Write deterministic tests that confirm insecure, corrected, and negative-case behavior.
- Propose suitable safeguards and ways to remediate identified weaknesses.
- Contribute to technical alignment sessions, colleague reviews, and quality checks.
Requirements
- At least 5+ years of professional work in application security, product security, or software engineering centered on security.
- Strong capability in secure code review and finding vulnerabilities at the application layer.
- Practical command of taint analysis, source-to-sink tracing, and SAST.
- Hands-on use of analysis products such as CodeQL or Semgrep.
- Direct experience developing LLM agents or RAG applications is required; relevant work may also include protecting these systems.
- Familiarity with LangChain, LlamaIndex, OpenAI or Anthropic SDKs, and MCP.
- Thorough knowledge of authorization concepts involving actors, identities, and trust boundaries.
- Understanding of access control and authorization for multi-tenant environments.
- Knowledge of OAuth, IAM, and identity management.
- Ability to reason about which information and actions are permitted, including constraints based on purpose or recipient.
- Understanding of controls governing access at document level.
- Production-grade development ability in Python and/or TypeScript.
- Ability to separate genuine SID and EA weaknesses from false positives and cases that do not constitute findings.
- Strong analytical capability and clear communication of technical security results.
How to apply
Submit an application through the employer’s official vacancy for this remote position.