Bengali Transcription & Annotation Specialist

Obsidian; Oslo; On-site; NOK 420,000 - 640,000; Full time

Benefits offered by this job

Regional-language data

Transcription and localization

Document digitization

Unicode encoding accuracy

Job summary

Mercor is building training data for Bengali documents, focusing on handwriting, dense column layouts, tables, and mixed scripts. You will annotate structure, bound regions, and assign reading order while ensuring faithful Bengali transcription.

You will also capture metadata and verify work through peer review to maintain high quality across hundreds of pages. We seek native Bengali speakers with exacting attention to diacritics, conjunct forms, and script fidelity.

Qualifications

• You are a native Bengali speaker with full command of the script, its diacritics and conjunct forms

Responsibilities

• Source documents: find publicly accessible Bengali PDFs with multimodal elements and record the source
• Annotate structure: bound every meaningful region and assign a type and reading order
• Record relationships: link regions to their parent figure or table
• Transcribe faithfully: reproduce Bengali text exactly, flag unreadable regions
• Capture page metadata: language, document type, source, dimensions, and flags for tables and handwriting
• Review a colleague's work: every task reviewed by a second Bengali expert

Job description

Mercor is building training data for Bengali documents, focusing on handwriting, dense column layouts, tables, and mixed scripts. You will annotate structure, bound regions, and assign reading order while ensuring faithful Bengali transcription.

You will also capture metadata and verify work through peer review to maintain high quality across hundreds of pages. We seek native Bengali speakers with exacting attention to diacritics, conjunct forms, and script fidelity.