Korean PDF Annotation Specialist | Structure Transcription

Obsidian; Madrid; Presencial; EUR 17.000 - 30.000; Jornada completa

Descripción de la vacante

Mercor is seeking a native Korean annotator to build training data for Korean documents, including newspapers, exams, and forms. You will bound and classify page regions, ensure precise Hangul transcription, and record page metadata. Collaboration with a second Korean expert ensures final quality.

The role emphasizes exactness, taxonomy consistency, and comfort with diverse layouts to enrich a multilingual corpus.

Formación

• Native Korean speaker with full command of Hangul and hanja where present.
• Experience in bilingual transcription, editorial work, or AI training data preferred.
• Accuracy is paramount; character-level correctness is required.
• Apply a consistent taxonomy across hundreds of pages.
• Comfortable with unfamiliar layouts: multi-column newspapers, exams, forms.

Responsabilidades

• Source publicly accessible Korean PDFs in assigned document types with multimodal elements.
• Annotate structure: bound regions with type and reading-order index.
• Record relationships linking regions to parent figures or tables.
• Transcribe text exactly in Hangul, flag non- legible regions.
• Capture page metadata: language, type, source, dimensions, and flags.
• Review a colleague's work; undergo second-expert verification.

Conocimientos

Korean language proficiency

Native Korean speaker

Attention to character-level accuracy

Bilingual transcription experience

Editorial/reviewer experience

Systematic taxonomy application

Adaptability to varied layouts

Descripción del empleo

Mercor is seeking a native Korean annotator to build training data for Korean documents, including newspapers, exams, and forms. You will bound and classify page regions, ensure precise Hangul transcription, and record page metadata. Collaboration with a second Korean expert ensures final quality.

The role emphasizes exactness, taxonomy consistency, and comfort with diverse layouts to enrich a multilingual corpus.