AI Adversarial Robustness Testing Engineer

Yovex Intelligence

Remote · $160k – $200k / year · Full-time · Posted 4 months ago

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

About the role

LLMs break in ways that are impossible to predict without adversarial testing. A model that performs flawlessly on benchmark evaluations can be jailbroken with a specific unicode sequence, manipulated into revealing system prompt contents through indirect injection, or induced into confident hallucination by a carefully constructed context. Enterprise clients who deploy LLMs in production face these risks and most of them lack the expertise to evaluate them systematically. We're building the adversarial testing and red teaming platform that enterprise AI teams use before deployment. The platform runs automated attack suites — jailbreaks, prompt injection vectors, adversarial robustness tests — and produces an exploitability report that security and AI teams can act on. You'll lead the attack suite development: researching published attack techniques, implementing them as automated test cases, discovering novel attacks through manual exploration, and building the evaluation framework that scores model robustness. Your work will directly inform deployment decisions at Fortune 500 companies.

Fully remote.

Responsibilities

Implement automated attack suites for jailbreaking, prompt injection, and adversarial robustness
Research and adapt published adversarial attack techniques into test cases
Discover novel attack vectors through manual red teaming and systematic exploration
Build exploitability scoring and risk classification for attack findings
Create detailed remediation guidance for common vulnerability patterns
Publish findings on novel attack techniques to the security research community

Requirements

5+ years in ML engineering or security research with LLM adversarial testing experience
Deep familiarity with published jailbreak and prompt injection techniques
Strong Python and PyTorch skills
Experience with red teaming, penetration testing, or security research methodology
Publications or public contributions in AI safety, security, or adversarial ML