[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"job-contract-ai-generalist-trainer-germany":3},{"Ques":4,"Slug":22,"Header":23,"job_category":39},{"title":5,"content":6},"Frequently Asked Questions",[7,10,13,16,19],{"A":8,"Q":9},"Yes. This is a fully remote, full-time role, and you must be based in Germany.","Is this role remote?",{"A":11,"Q":12},"You will perform large language model evaluation, including RLHF-style ranking of model outputs, prompt evaluation, QA evaluation, validation against guidelines, and writing clear rationales to support training data quality and model performance improvement.","What tasks will I do?",{"A":14,"Q":15},"AI experience is helpful but not required. We value strong analytical skills, attention to detail, and the ability to follow annotation guidelines compliance and provide consistent evaluations.","Do I need AI experience?",{"A":17,"Q":18},"You must be fluent in both German and English, since evaluations and rationales will be completed in both languages.","What languages are required?",{"A":20,"Q":21},"The work can span general knowledge, reasoning, writing quality, instruction following, and content safety labeling, depending on project needs within AI\u002FLLM workflows.","What domains are covered?","contract-ai-generalist-trainer-germany",{"desc":24,"title":25,"content":26},"Rexzone is hiring Germany-based bilingual (English\u002FGerman) AI Generalist Trainers to support RLHF and large language model evaluation by assessing, ranking, and QA-checking model outputs. You will apply annotation guidelines compliance to improve training data quality, write clear rationales, and help drive model performance improvement across real-world AI\u002FLLM workflows in a fully remote, full-time role.","Germany-Based English & German AI Generalist Trainer 2026 May",[27,30,33,36],{"h2":28,"desc":29},"About the Role","As an AI Generalist Trainer at Rexzone, you will evaluate and improve AI systems by reviewing model-generated responses in English and German. Your work centers on RLHF-style ranking, prompt evaluation, QA evaluation, and validation against policy and annotation guidelines compliance. You will document reasoning, flag content safety issues, and ensure training data quality for reliable model performance improvement.",{"h2":31,"desc":32},"Responsibilities","• Perform large language model evaluation by reviewing prompts and model outputs in English and German\n• Rank and compare multiple model responses using RLHF-aligned criteria\n• Write concise, evidence-based rationales that capture reasoning and evaluation judgments\n• Execute QA evaluation to validate annotation accuracy, consistency, and annotation guidelines compliance\n• Label and categorize data for data labeling workflows, including content safety labeling when required\n• Validate edge cases, detect policy violations, and escalate ambiguous items with clear notes\n• Track common failure modes and propose rubric improvements that enhance training data quality\n• Maintain high throughput while meeting accuracy targets and quality benchmarks",{"h2":34,"desc":35},"Basic Qualifications","• Must be based in Germany and able to work remotely from Germany\n• Fluent in German and English (reading, writing, and comprehension)\n• Strong analytical skills with the ability to compare outputs and justify rankings\n• Excellent attention to detail and consistency in applying rubrics and guidelines\n• Comfortable working with structured tasks, feedback loops, and quality audits\n• Able to explain reasoning clearly and objectively in written rationales",{"h2":37,"desc":38},"Preferred Qualifications","• Prior experience with AI data labeling, LLM evaluation, prompt evaluation, or QA evaluation\n• Familiarity with RLHF concepts, ranking tasks, and training data quality best practices\n• Experience following annotation guidelines and contributing to rubric refinement\n• Knowledge of content safety labeling concepts (toxicity, policy compliance, sensitive content)\n• Self-driven, reliable, and able to manage time effectively in a remote environment","AI Data Operations"]