Remote AI Safety Tester — Adversarial Prompt Designer

Alignerr Corp.; Argentina; Presencial; ARS 41.988.000 - 94.474.000; Jornada completa

Ventajas ofrecidas por este puesto de trabajo

Fully remote

Flexible hours

Contract freelance

Descripción de la vacante

Alignerr is seeking AI Red Team Testers to probe and stress-test cutting-edge AI systems, thinking like adversaries to expose weaknesses before they cause real-world harm. This is a fully remote hourly contract, open to anyone who enjoys puzzles and critical thinking.

You design adversarial prompts, elicit unsafe outputs, and document failures with reproducible steps, collaborating asynchronously with AI safety teams. No cybersecurity background required.

Formación

• Creative, curious thinker who enjoys puzzles and analysis.
• Comfortable thinking adversarially and asking what could go wrong?
• Detail-oriented with clear, thorough documentation.
• Strong written English communication.

Responsabilidades

• Design adversarial prompts and scenarios to expose AI weaknesses
• Attempt to elicit unsafe or biased outputs
• Document failures with reproducible steps
• Rate and categorize issues by severity
• Collaborate asynchronously with safety and research teams
• Explore edge cases across topics, tones, and scenarios

Conocimientos

Creative thinking

Curious

Puzzle solving

Attention to detail

English writing

Independent

Adversarial mindset

Descripción del empleo

Alignerr is seeking AI Red Team Testers to probe and stress-test cutting-edge AI systems, thinking like adversaries to expose weaknesses before they cause real-world harm. This is a fully remote hourly contract, open to anyone who enjoys puzzles and critical thinking.

You design adversarial prompts, elicit unsafe outputs, and document failures with reproducible steps, collaborating asynchronously with AI safety teams. No cybersecurity background required.