Bengali PDF Annotation & Transcription Expert | $12.68/hr | Remote (India, US, Canada, Western Europe)
Mercor is seeking native Bengali-speaking annotation and transcription specialists to build high-quality document AI training data. This role focuses on structurally mapping and faithfully transcribing real Bengali PDF documents — helping language models better understand scripts and layouts they currently handle poorly.
About the Project
This initiative creates training data for document understanding models across underrepresented Indic scripts, including Bengali. You will work with real, publicly available PDF pages — newspapers, textbooks, exam papers, forms, menus, flyers, and more — annotating their structure and transcribing all text content in Bengali script with character-level precision.
What You'll Do
- Source publicly accessible Bengali PDFs in assigned document types that include at least one multimodal element (images, tables, diagrams, or handwriting)
- Annotate every meaningful page region — titles, headings, paragraphs, lists, tables, figures, captions, formulas, and answer fields — assigning component types and reading-order indices
- Record parent-child relationships between regions and their associated figures or tables
- Transcribe all text exactly as it appears in Bengali script, including handwritten content, and flag any illegible regions
- Capture page-level metadata: language, document type, source URL, dimensions, and content flags
- Perform peer review of a colleague's completed tasks as you gain experience
Who You Are
- Native Bengali speaker with full command of Bengali script, diacritics, and conjunct character forms
- Experienced in document-related work: annotation, transcription, translation, localization, proofreading, or regional-language data review
- Highly precise — character-level accuracy is the top priority; a single incorrect diacritic is considered a defect
- Systematic and consistent in applying taxonomies across large volumes of varied documents
- Comfortable navigating complex layouts such as multi-column newspapers, exam papers, and handwritten forms
Nice-to-Have Skills
- Experience with regional-language AI data: annotation, labeling, bilingual evaluation, or dataset QA
- Background in MTPE, subtitling, OCR correction, or bilingual post-editing
- Document production experience: typesetting, copy-editing, or digitization of Bengali-language materials
- Knowledge of Unicode normalization, Bengali input methods, and numeral/character encoding accuracy
What Success Looks Like
- All meaningful page regions are captured, correctly bounded, and accurately typed
- Reading order faithfully reflects how the page is read, including across columns
- Transcriptions match the source character-for-character in Bengali script (no transliteration)
- Tasks pass second-expert review on the first submission
- Sourced documents add genuine layout diversity to the corpus
This is a remote contractor role open to contributors based in India, the United States, Canada, or Western Europe. All work is performed in Bengali.
