Korean PDF Annotation Specialist — Precise Multimodal

Mercor; Greater London; On-site; GBP 24,000 - 36,000; Full time

Job summary

Mercor is seeking a native Korean speaker for data annotation and transcription work to build Korean datasets for document AI training. You will bound regions, ensure reading order, and transcribe Hangul with exactness.

Ideal candidates have experience in bilingual transcription, editorial work, or AI training data, with a meticulous eye for detail and consistent taxonomy across pages.

Qualifications

• Native Korean speaker with full command of Hangul.
• Experience in bilingual transcription, editorial work, or AI training data is preferred.
• Accuracy is critical; character-level precision is essential.

Responsibilities

• Source Korean PDFs and annotate structure: title, headings, paragraphs, lists, tables, figures, captions, formulas.
• Bound regions and assign component types and reading-order indices.
• Record relationships between regions and parent tables/figures.
• Transcribe all text exactly as Hangul, flag illegible regions.
• Capture metadata: language, document type, source, page dimensions, handwriting flags.
• Participate in peer review by a second Korean expert.

Skills

Korean fluency

Bilingual transcription

Attention to detail

Job description

Mercor is seeking a native Korean speaker for data annotation and transcription work to build Korean datasets for document AI training. You will bound regions, ensure reading order, and transcribe Hangul with exactness.

Ideal candidates have experience in bilingual transcription, editorial work, or AI training data, with a meticulous eye for detail and consistent taxonomy across pages.