Korean PDF Annotation Specialist for AI Training Data

Mercor; København; On-site; DKK 250,000 - 420,000; Full time

Job summary

Mercor is seeking a meticulous Korean annotator for building training data for document AI. You will identify and bound every meaningful region on Korean PDFs with multimodal content—tables, diagrams, handwriting—and assign reading order and types.

You will transcribe text in Hangul, capture metadata, and ensure accurate, reviewer-verified outputs. Applicants should be native Korean speakers with experience in transcription or editorial work and a keen eye for detail.

Qualifications

• Native Korean speaker with full command of Hangul and hanja where it appears.
• Experience in bilingual transcription, translation, editorial work, or AI training data preferred.
• Extremely exact, character-level accuracy matters more than speed.

Responsibilities

• Source publicly accessible Korean PDFs in assigned document types with multimodal elements.
• Annotate page structure: bound regions, assign types, and reading-order indices.
• Record relationships by linking regions to their parent figures/tables.
• Transcribe faithfully in Hangul, flag illegible regions.
• Capture page metadata: language, document type, page dimensions, and flags for tables, formulas, handwriting.
• Review a colleague's work through second-expert checks.

Skills

Korean language fluency

Attention to detail

Transcription/annotation

Quality review

Documentation experience

Job description

Mercor is seeking a meticulous Korean annotator for building training data for document AI. You will identify and bound every meaningful region on Korean PDFs with multimodal content—tables, diagrams, handwriting—and assign reading order and types.

You will transcribe text in Hangul, capture metadata, and ensure accurate, reviewer-verified outputs. Applicants should be native Korean speakers with experience in transcription or editorial work and a keen eye for detail.