Bengali Document Annotation Specialist

Obsidian; Lisboa; Presencial; EUR 13 000 - 20 000; Tempo integral

Resumo da oferta

Mercor seeks native Bengali speakers to annotate and transcribe Bengali documents with exact script fidelity. You will identify regions, assign types, and ensure reading order for complex layouts.

You will work on publicly accessible Bengali PDFs, handling handwriting, tables, and diagrams, and participate in peer reviews to guarantee high quality and consistency.

Collaboration and meticulous verification are essential to meet project standards and improve dataset diversity for AI training.

Qualificações

• Native Bengali speaker with full command of Bengali script and orthography.
• Experience with document annotation, transcription, localization, or similar data-labeling tasks.
• Excellent accuracy in character-level transcription and careful handling of complex layouts.

Responsabilidades

• Source Bengali PDFs with multimodal content for annotation tasks.
• Annotate page structure: identify regions (title, heading, paragraph, table, figure, caption, etc.) and assign types and reading-order indices.
• Record relationships by linking regions to their parent figure/table.
• Transcribe faithfully in Bengali script, flagting illegible regions.
• Capture page metadata: language, document type, source, dimensions, and flags for tables/form/handwriting.
• Review a colleague's work to ensure high-quality, accurate results.

Conhecimentos

Bengali fluency

Attention to detail

Annotation experience

Descrição da oferta de emprego

Mercor seeks native Bengali speakers to annotate and transcribe Bengali documents with exact script fidelity. You will identify regions, assign types, and ensure reading order for complex layouts.

You will work on publicly accessible Bengali PDFs, handling handwriting, tables, and diagrams, and participate in peer reviews to guarantee high quality and consistency.

Collaboration and meticulous verification are essential to meet project standards and improve dataset diversity for AI training.