AI Trainer
TextRLHF
As an AI Trainer at Outlier AI, I evaluated and ranked AI-generated responses as part of RLHF workflows. I authored Japanese training data to enhance the LLM’s capabilities and assessed the factual accuracy, reasoning, and cultural appropriateness of outputs. My work contributed directly to improving generative model quality, involving structured, repeated evaluations of model behavior.
• Conducted rigorous AI output rating through structured RLHF pipelines
• Developed Japanese textual prompts and responses for LLM training
• Reviewed cultural appropriateness and bias in language model completions
• Regularly collaborated remotely with cross-lingual and international team
As an AI Trainer at Outlier AI, I evaluated and ranked AI-generated responses as part of RLHF workflows. I authored Japanese training data to enhance the LLM’s capabilities and assessed the factual accuracy, reasoning, and cultural appropriateness of outputs. My work contributed directly to improving generative model quality, involving structured, repeated evaluations of model behavior. • Conducted rigorous AI output rating through structured RLHF pipelines • Developed Japanese textual prompts and responses for LLM training • Reviewed cultural appropriateness and bias in language model completions • Regularly collaborated remotely with cross-lingual and international team
2024 - Present