Image Caption Review - Japanese (Bilingual)

About this role

Verify Japanese text that appears inside images against the caption written about it. You will check captions produced by an AI model against the image they describe, focusing on text rendered in the image itself — signage, labels, packaging, screenshots. Where the caption misreads, mistranslates or omits that text, you log the defect with the exact wording quoted. Your written reviews are in English; the Japanese is what you read in the image. Requirements: native-level Japanese and fluent English, bachelor's degree, 2+ years professional experience, and the patience to compare text character by character.

Skills & domains

• ai-training
• rlhf
• sme
• annotation