Germany-Based English & German AI Generalist Trainer 2026 May - #2727359

About The Role

As a Germany-based English & German AI Generalist Trainer at Rex.zone, you will evaluate and improve AI systems by reviewing model-generated outputs in real-world scenarios. Your work supports RLHF, large language model evaluation, and consistent training data quality through careful labeling, prompt evaluation, and annotation guidelines compliance.

Key Responsibilities

• Evaluate and rank model-generated responses against defined rubrics and project goals
• Perform QA evaluation by validating labels, rationales, and final outputs for accuracy and consistency
• Write concise reasoning and justifications in German and English to support RLHF datasets
• Apply annotation guidelines compliance and flag ambiguity or guideline gaps for resolution
• Execute prompt evaluation and large language model evaluation across multiple task types
• Conduct validation checks for content safety labeling and policy adherence
• Identify recurring failure modes and suggest improvements that increase training data quality and model performance
  
  

Basic Qualifications

• Must be based in Germany (remote role) and authorized to work as applicable
• Fluent in German and English (reading and writing) with strong grammar and style
• Strong analytical skills, attention to detail, and ability to follow multi-step rubrics
• Ability to document clear, consistent rationales
• Reliable internet connection and ability to meet quality and productivity targets
  
  

Preferred Qualifications

• Experience with data labeling, QA evaluation, or annotation projects
• Familiarity with LLMs, RLHF concepts, and evaluation workflows
• Background in linguistics, translation, writing, journalism, customer support, or technical documentation
  
  

How To Apply

Apply with a current CV/resume and a brief note confirming Germany location and German/English fluency.

Qualified candidates may be asked to complete an online assessment focused on evaluation, ranking, reasoning, and training data quality.