Korean Transcription Specialist for AI Training Data

Obsidian; Oslo; Remote; NOK 450,000 - 550,000; Full time

Job summary

Obsidian in Oslo seeks meticulous Korean-language annotators to build training data for Korean, Japanese and Indic scripts. You will process real PDFs to produce a complete structural map and a faithful Hangul transcription.

You must be a native Korean speaker with strong accuracy, and you will work under a two-expert review process to ensure quality across hundreds of pages with multi-column layouts and varied content.

Qualifications

• Native Korean speaker with Hangul mastery and hanja familiarity where present.
• Experience in transcription, translation, or AI training data is preferred.
• Very precise, systematic and capable of handling complex layouts.
• Respect for typography, reading order and annotation taxonomy.

Responsibilities

• Open and check a task: pages are provided, confirm Korean content and legibility.
• Annotate structure: bound every meaningful region and assign a type and reading-order index.
• Record relationships: link regions to the figure or table they belong to via a parent identifier.
• Transcribe faithfully: reproduce all text exactly in Hangul, flag illegible regions.
• Capture page metadata: language, document type, source, dimensions, and flags for tables, formulas and handwriting.
• Review a colleague's work: second Korean expert reviews each task.

Skills

Native Korean speaker

Attention to detail

Character-level accuracy

Editorial background

Bilingual transcription experience

Job description

Obsidian in Oslo seeks meticulous Korean-language annotators to build training data for Korean, Japanese and Indic scripts. You will process real PDFs to produce a complete structural map and a faithful Hangul transcription.

You must be a native Korean speaker with strong accuracy, and you will work under a two-expert review process to ensure quality across hundreds of pages with multi-column layouts and varied content.