LLM Jailbreak Red Teamer

Claru; United States; Remote; USD 76,000 - 138,000; Part time

Job summary

AI Trainer Jobs is recruiting for a role focused on adversarial prompt research. You will systematically probe large language models to identify inputs that bypass safety guidelines, documenting prompts, model outputs, and why results violate policy.

Sessions are scoped and time-boxed, with structured logging and escalation paths. The role emphasizes methodical attack vectors, with a preference for candidates who understand reports, taxonomy, and reproducible testing.

Qualifications

• Experience identifying and classifying harmful prompts.
• Ability to document attack vectors with reproducible steps.
• Understanding of policy violations and severity assessment.

Responsibilities

• Plan and execute adversarial prompt experiments within defined harm categories.
• Log prompts, model outputs, and severity ratings for each attempt.
• Document findings and contribute to a findings database.

Skills

Adversarial prompt crafting

Harm taxonomy classification

Multi-turn attack strategies

Structured vulnerability documentation

Policy violation severity rating

Jailbreak technique knowledge

Job description

Pay: $55–100/hour

You'll systematically probe large language models to find inputs that cause them to bypass safety guidelines and produce harmful, disallowed, or policy-violating outputs. Your work involves crafting adversarial prompts across defined harm categories — including but not limited to dangerous instructions, deceptive content, and identity manipulation — then documenting exactly what worked, what the model produced, and why the response constitutes a policy violation. You won't be testing randomly; each session is scoped to specific harm categories and model behaviors as directed by the research team.

The workflow is structured: you'll receive a target model, a harm taxonomy, and a session goal. You'll then attempt multiple attack vectors per category — direct prompting, role-play framing, multi-turn escalation, encoding tricks, and others — logging each attempt with the full prompt, model response, and a severity rating. Successful jailbreaks get promoted to a verified findings database; failed attempts are still valuable and should be logged with notes on why they failed. Sessions typically run 2–3 hours with a defined scope.

The strongest candidates are people who think like adversaries — researchers, security professionals, former CTF competitors, or anyone who has spent time stress-testing systems. You should be methodical enough to document findings reproducibly and creative enough to try non-obvious attack angles. Familiarity with existing jailbreak literature (DAN prompts, prompt injection, etc.) is assumed. You won't be writing exploit code, but you will be thinking in the same mode.

Skills & requirements

• Adversarial prompt crafting
• Harm taxonomy classification
• Multi-turn attack strategies
• Structured vulnerability documentation
• Policy violation severity rating
• Jailbreak technique knowledge

Category: Red Teaming