Cybersecurity Experts for AI Training

About the role

Gramian Consultancy is seeking experienced cybersecurity professionals to design and implement realistic security‑focused coding scenarios for evaluating frontier AI systems. The role blends offensive security, application security, and AI evaluation to create reproducible benchmarks that test AI understanding of software vulnerabilities.

Key responsibilities

• Design multi‑component vulnerable applications and challenge environments in Go, Python, Node.js, or Rust.
• Develop exploit chains that combine multiple vulnerability categories and attack vectors.
• Build deterministic evaluation environments using Docker and automated validation tooling.
• Create security‑focused test cases and verification logic for vulnerability detection and remediation.
• Review AI‑generated outputs to spot gaps in reasoning, security understanding, or exploit detection.
• Model real‑world vulnerability classes inspired by CVEs, bug bounty findings, and production incidents.
• Ensure tasks remain scalable, reproducible, and resistant to contamination from public datasets.
• Collaborate with cross‑functional AI evaluation and benchmarking teams.

Required profile

• 4+ years of experience in cybersecurity, application security, vulnerability research, or offensive security.
• Hands‑on experience with vulnerability discovery, exploit development, secure code review, or patch validation.
• Strong understanding of web security, authentication, sessions, OAuth, JWT, SSRF, injection attacks, and access‑control vulnerabilities.

Required skills

• Go, Python, Node.js, Rust (at least two)
• Docker and containerisation
• SAST, fuzzers, IAST or similar security testing frameworks
• Cryptographic vulnerability analysis
• Filesystem attacks and privilege‑escalation techniques