Bengali Transcription & Page-Structure Annotation Pro

Obsidian; Dublin; On-site; EUR 32,000 - 44,000; Full time

Job summary

Obsidian seeks a meticulous Bengali annotation specialist to build training data from real Bengali documents. You will identify, bound, and classify page regions such as titles, headings, paragraphs, tables, and figures, ensuring precise reading order and faithful transcription.

The role emphasizes character-level accuracy, handling diverse layouts, and collaboration with senior annotators for end-to-end quality review.

Qualifications

• Native speaker of Bengali with full command of Bengali script and orthography.
• Experience in annotation, transcription, localization, or related data reviews is desirable.
• Character-level accuracy is essential; even a single diacritic mistake is a defect.

Responsibilities

• Source publicly accessible Bengali PDFs containing at least one multimodal element.
• Annotate and bound every meaningful region with type and reading-order index.
• Record relationships by linking regions to their parent figures/tables.
• Transcribe faithfully in Bengali script, flag unreadable regions.
• Capture page metadata: language, document type, source, dimensions, and handwriting/table flags.
• Review work with a second Bengali expert to ensure accuracy.

Skills

Bengali language proficiency

Attention to detail

Systematic annotation

Job description

Obsidian seeks a meticulous Bengali annotation specialist to build training data from real Bengali documents. You will identify, bound, and classify page regions such as titles, headings, paragraphs, tables, and figures, ensuring precise reading order and faithful transcription.

The role emphasizes character-level accuracy, handling diverse layouts, and collaboration with senior annotators for end-to-end quality review.