Security is one of the areas where AI models are most confidently wrong — and where wrong answers matter most. Labs pay cybersecurity professionals to evaluate model output on vulnerability analysis, secure coding, threat modeling, and incident response, and to red-team the models themselves. If you work in security or IT, that's remote, flexible work at $50–$150/hr — and your adversarial instincts are exactly what the job rewards.
What security AI-training work involves
- Vulnerability & secure-code review — judge whether the model correctly identifies (or introduces) security flaws in code and configs.
- Threat modeling & reasoning evaluation — score the model's analysis of attack surfaces, mitigations, and risk.
- Red-teaming the model — probe for unsafe outputs, jailbreaks, and confidently-wrong security advice. Pentesters excel here.
- Reference-answer generation — write the gold-standard security analysis the model trains on (top-paid).
Roles open now
- Cybersecurity Expert (AfterQuery) — $50–$100/hr
- Software engineering / secure-code evaluation across platforms — $50–$150/hr for senior backgrounds
Browse them on our AfterQuery listings page or filter the board by coding / software.
Which platforms hire security & IT pros
- AfterQuery — the dedicated Cybersecurity Expert role plus engineering evaluation work.
- Turing — engineering-heavy, strong for security-adjacent code review; see our software engineers guide.
- Mercor — engine-matched frontier-lab work for senior security engineers.
How to position a security background
- Name the specialty. AppSec, offensive security, cloud security, GRC, DFIR — specificity routes you to the right role and rate.
- List certs and proof. OSCP, CISSP, CEH, CVEs you've found, CTF results, bug-bounty history — all verifiable signals.
- Lead with hands-on experience. Actual pentesting or secure-code review beats theory for the red-teaming work.
Interview tactics are in our interview guide.
A note on ethics & scope
This work is defensive by design — you're helping models give safer, more accurate security guidance and refuse harmful requests. It's legitimate contract work, paid weekly/biweekly like any other AI training role. As always, no real platform charges a fee to start — see the scam guide.
Get started
Apply to the Cybersecurity Expert role on the AfterQuery page, add Turing and Mercor for senior engineering work, and read the complete AI training jobs guide for the full landscape.