Japanese Document Structure & Transcription Specialist

ixolabs.ai; France; Hybride; EUR 42 000 - 50 000; Plein temps

Résumé du poste

IXO seeks contributors to create a faithful, human-authored transcription and structural representation of Japanese PDF pages, including difficult layouts, handwriting, mixed scripts and tables.

You will bound regions, assign reading order, link figures/tables, and verify language accuracy with a focus on precision. This is a remote, hourly engagement with clear deliverables and quality checks.

Qualifications

• Native Japanese required with knowledge of kanji, hiragana, katakana.
• Attention to character-level accuracy over speed.
• Experience with transcription, editing, or AI data work preferred.

Responsabilités

• Inspect pages for language, legibility, content and personal data; flag unsuitable material.
• Bound all meaningful regions (headings, paragraphs, tables, figures, captions, formulas, questions, answers).
• Assign reading-order positions and cross-link regions to figures or tables.
• Transcribe text faithfully and flag unreadable passages.
• Record page metadata, language and flags for handwriting or tables.

Description du poste

Your work at IXO

Create a human-checked structural and textual representation of Japanese PDF pages. The work covers difficult layouts, handwriting, mixed scripts, tables and diagrams, with exact transcription in the original script.

Responsibilities

• Inspect the supplied page for the correct language, legibility, meaningful content and absence of personal details; flag unsuitable material before annotation.
• Bound every meaningful region, including headings, paragraphs, lists, tables, figures, captions, formulas, questions and answer fields.
• Assign component types and reading-order positions, including across columns.
• Link subordinate regions to the relevant parent figure or table.
• Transcribe every text region faithfully and flag unreadable passages instead of guessing.
• Record page dimensions, document type, source, language and flags for handwriting, formulas or tables.
• Review another contributor’s complete page when assigned; each task receives a second expert check.

Background and tools

Native Japanese and full command of kanji, hiragana, katakana, furigana and variant forms, including vertical text and full-/half-width differences are required. Bring experience in transcription, translation, editing or AI data work, preferably including review. You must apply a taxonomy consistently and prioritize character-level accuracy over speed.

OCR correction, localization, typesetting, proofreading, digitization, Unicode normalization and language input methods are useful specialties. All structure and text work must be human-authored; parsing-model output is not a substitute.

Deliverables and pay

Deliverables are the complete region map, relationships, reading order, page metadata and script-faithful text. Pages are supplied for the task; contributors do not need to collect private documents. IXO will confirm the scope and schedule before work begins.

Pay is $38.58/hr in USD. Work is remote.