Get a reply from this employer
Your resume and cover letter, tailored to exactly what they’re hiring for.
GraceMark Solutions is seeking a Cybersecurity SME specializing in AI Red Teaming to probe AI models and coding agents for cybersecurity risks in isolated sandbox environments. You will iteratively test exploit scenarios, assess safety, and mentor junior red teamers and annotators.
You will map findings to MITRE ATLAS, OWASP Top 10 for LLM Applications, and NIST AI RMF, while developing realistic attack datasets and documenting reproducible steps for researchers, engineers, and executives.
The Cybersecurity SME, AI Red Team acts as an offensive security practitioner who probes AI models and coding agents for cybersecurity risks, exploit generation capabilities, and agentic vulnerabilities. Working inside isolated sandbox environments, you will iteratively test whether models can generate functional exploits, malware, or attack tooling, evaluate findings against strict safety specifications, map risks to established AI/cyber frameworks, and mentor junior red teamers and annotators.
Model & Coding Agent Red-Teaming: Iteratively probing AI coding agents and frontier LLMs in sandboxed environments to evaluate whether models can generate functional offensive tooling, exploits, and malware.
Capability Uplift & Safety Evaluation: Assessing model outputs against established safety specifications, determining violation severity, and refining criteria for unsafe cyber capability uplift.
Exploit & Malware Verification: Evaluating model performance in vulnerability discovery, exploit development, and malware analysis, verifying in sandboxes whether AI-generated tools actually execute as intended.
Agent & Harness Security Testing: Investigating agent-specific attack surfaces, including prompt injection, tool misuse, and privilege escalation as evaluation expands into complex agent harnesses.
Dataset Creation & Quality Review: Developing realistic attack scenarios and evaluation datasets, and reviewing expert-written prompts and model outputs for technical accuracy.
Framework Mapping & Reporting: Documenting reproduction steps, severity ratings, and mitigations, and mapping findings to MITRE ATLAS, OWASP Top 10 for LLM Applications, and NIST AI RMF.
Training & Mentorship: Training and mentoring red teamers and annotators on offensive security concepts while staying current on emerging AI attack techniques.
Hands-On Offensive Security Experience: 2+ years in penetration testing, red teaming, vulnerability research, or exploit development.
Practical Attack & Exploit Knowledge: Direct experience with exploit development, malware behavior, post-exploitation techniques, and web or cloud attacks.
Scripting & Code Validation: Strong scripting proficiency (preferably Python) with the ability to read, debug, and safely execute/verify AI-generated offensive code in a sandbox.
AI / LLM Security Awareness: Working knowledge of LLMs, AI coding agents, and common attack vectors targeting them (e.g., jailbreaks, prompt injection, tool misuse).
Clear Technical Writing: Strong written communication skills to document reproduction steps and explain complex security risks to researchers, engineers, and executives.
Content Resilience & OpSec: Ability to handle sensitive or adversarial cyber content responsibly while strictly adhering to confidentiality and information-handling protocols.
Preferred Qualifications: Certifications or proven track record via OSCP, OSEP, GPEN, published CVEs, CTF placements, bug bounties, AI/ML security research, defense/intelligence background, malware reverse engineering, or cloud security (AWS, GCP, Azure).