Norwegian AI Trainer – Remote (Hourly)

About the role

We are looking for a Norwegian‑speaking AI Trainer to evaluate large language model responses for personalization quality. This remote, hourly‑based position involves testing multi‑turn prompts and ensuring the model delivers accurate, helpful, and context‑aware answers.

Key responsibilities

• Assess AI model outputs for grounding, integration, helpfulness, and personalization accuracy.
• Design and execute multi‑turn prompts that reflect personal context to probe model capabilities.
• Identify hallucinations, incorrect personalization, and weak inferences in responses.
• Perform side‑by‑side comparisons of model outputs to rank quality and effectiveness.
• Document clear rationales for each evaluation and maintain structured records.
• Extract and verify debug information to confirm correct data source usage.
• Ensure strict data hygiene and accurate documentation throughout the evaluation process.
• Collaborate with cross‑functional teams to improve model performance.

Required profile

• Native‑level proficiency in Norwegian with excellent reading and writing abilities.
• Experience in data annotation, AI evaluation, content moderation, or related analytical fields.
• Strong analytical thinking and ability to evaluate nuanced AI responses.
• Capability to design creative, multi‑turn prompts based on personal context.
• Understanding of personalization concepts and ability to spot forced or incorrect personalization.
• High attention to detail and excellent written communication skills.
• Self‑motivated and able to work independently in a remote setting.
• Availability for full‑time work with at least 4 hours overlap with Pacific Standard Time.

Required skills

What we offer

• Hourly compensation ranging from $11 to $15 USD.
• Fully remote work environment.
• Flexible schedule with a 30–40 hour weekly commitment.