Talent.com
Cette offre d'emploi n'est pas disponible dans votre pays.
AI Adversarial Tester / Prompt Writer

AI Adversarial Tester / Prompt Writer

TaskUsBelgium, Belgium
Il y a 6 jours
Description de poste

About Us – TaskUs & TaskVerse

TaskUs provides digital outsourcing and next-generation customer experience for some of the world’s most innovative tech companies. We help companies build safer, smarter AI by delivering high-quality human feedback, content moderation, and data annotation services at scale.

TaskVerse is our freelance platform that connects global talent with flexible, project-based work. Taskers contribute to AI development through tasks like prompt writing, data labeling, transcription, and adversarial testing—helping improve the safety, fairness, and performance of emerging technologies.

About the Role

We’re looking for a freelance German-speaking AI Adversarial Tester / Prompt Writer to help test and improve large language models (LLMs) by crafting creative, edge-case, or challenging prompts.

This is a high-impact opportunity for individuals passionate about generative AI, critical thinking, and linguistic precision. You’ll be working at the frontier of AI safety and alignment—designing tests that push the boundaries of AI behavior.

Key Responsibilities

  • Design adversarial prompts to uncover weaknesses, safety risks, or edge cases in LLM behavior
  • Write clear, testable, and structured instructions to evaluate AI systems’ performance
  • Collaborate with a remote team to review outputs and escalate flagged issues
  • Generate diverse prompt types across multiple domains (e.g., ethical dilemmas, policy, humor, reasoning)
  • Work with internal guidelines to classify, label, or improve AI-generated content

Required Skills & Qualifications

You're a strong fit if you have experience in :

  • Working with LLMs (OpenAI, Anthropic, Cohere, Mistral, etc.)
  • Adversarial testing, red-teaming, or QA for AI systems
  • Writing high-quality prompts or structured inputs for generative models
  • Analytical thinking and debugging complex or ambiguous behavior in AI
  • Working independently across time zones and communicating asynchronously
  • Comfortable using Google Workspace tools (Docs, Sheets, Drive) for prompt writing, collaboration, and task submission
  • Fluency in English and German is essential
  • Preferred Qualifications

  • Previous experience in prompt writing for LLMs or other generative AI systems
  • Background in remote freelance work or global virtual teams
  • Interest in AI ethics, safety, or language fairness in generative systems
  • To apply, please submit your CV along with a brief explanation of your relevant experience and skills. As part of our screening, shortlisted candidates will complete a test to demonstrate prompt writing, localization, and basic understanding of AI safety concepts.

    Successful candidates will join our TaskVerse community and gain access to additional opportunities in the AI space.

    Important Note

    This role may involve testing AI systems with prompts that include sensitive or potentially challenging content. Candidates should be comfortable engaging with such material in a professional and constructive manner.

    Créer une alerte emploi pour cette recherche

    Tester • Belgium, Belgium