PDF Annotation & Transcription Expert – Malayalam

Contract · Remote · India

Annual Compensation Equivalent: $26,374.40/year
Hourly Rate: $12.68/hour

Fluent Language Skills Required: Malayalam

About the Role

This remote contract role is for native Malayalam speakers who can accurately annotate and transcribe real-world PDF documents to support the development of document understanding and vision-language AI systems.

You will create detailed structural annotations and faithful Malayalam transcriptions from publicly available documents. The project focuses on document types and layouts that are challenging for AI systems, including handwriting, dense multi-column pages, tables, diagrams, examinations, newspapers, forms, flyers, manuals, menus, brochures, notices and worksheets.

All annotation and transcription work is completed manually. Accuracy, consistency and faithful representation of the source document are critical.

What You'll Do

• Review assigned PDF pages before beginning work to confirm they are in Malayalam, legible, contain meaningful content and do not include personal information

• Identify and bound every meaningful region on each page, including titles, headings, paragraphs, lists, tables, figures, diagrams, captions, formulas, questions and answer fields

• Assign the appropriate component type and reading-order index to each annotated region

• Record relationships between regions and the figures or tables they belong to using parent component identifiers

• Transcribe all text exactly as it appears in the original Malayalam script, including handwritten content

• Flag text or regions that cannot be read reliably

• Capture page-level metadata such as language, document type, source and page dimensions

• Mark relevant document features such as tables, formulas and handwriting

• Review completed work from other Malayalam annotators for accuracy and consistency

Requirements

• Native-level Malayalam fluency with full command of Malayalam script and orthography

• Strong understanding of Malayalam diacritics, conjunct forms and character variations

• Experience working with documents through annotation, transcription, translation, localization, subtitling, proofreading, journalism or regional-language data review

• Exceptional attention to character-level accuracy

• Ability to apply a defined annotation taxonomy consistently across large volumes of documents

• Comfort working with unfamiliar and complex layouts, including newspapers, examination papers, multi-column documents and handwritten forms

• Ability to transcribe Malayalam directly in the original script rather than using transliteration

Preferred Qualifications

Experience in any of the following is helpful:

• Regional-language AI data annotation, labeling, grading or bilingual evaluation

• MTPE, subtitling, bilingual quality assurance or OCR correction

• Malayalam-language typesetting, copy-editing, proofreading or document digitization

• Unicode normalization and Malayalam input methods

• Malayalam numeral and character-form accuracy

Quality Expectations

Successful annotators consistently:

• Capture every meaningful region on a page

• Apply accurate boundaries and component types

• Assign reading order according to how the page should actually be read, including across columns

• Produce character-for-character Malayalam transcriptions

• Maintain consistent annotation decisions across documents

• Pass second-expert review with minimal corrections

• Identify unsuitable or illegible pages before spending significant time annotating them

Compensation & Engagement

• Hourly rate: $12.68/hour

• Annual compensation equivalent: $26,374.40 based on 40 hours per week, 52 weeks per year

• Actual earnings depend on hours worked and available project work

• Independent contractor engagement

• Fully remote

• Flexible schedule

• Projects may be extended, shortened or concluded early depending on project requirements and performance

Payment & Other Terms

• Payments are made weekly through Stripe or Wise based on services rendered

• The work does not require access to confidential or proprietary information belonging to an employer, client or institution

• H-1B and STEM OPT candidates are not currently supported