Malayalam Document Annotator and Transcriptionist for AI Training

SaidGig; New York (NY); Remote; USD 14,000 - 15,000; Part time

Benefits offered by this job

Remote work

Job summary

SaidGig is seeking skilled Malayalam annotators to build high-quality training data from real Malayalam PDFs. You will map page structure, transcribe every text region in the original script, and handle complex layouts such as multi-column pages and forms.

Ideal candidates are native Malayalam speakers with strong attention to character-level accuracy and experience with document annotation. This is a remote, hourly role with flexible hours, supporting a growing AI data project.

Qualifications

• Native fluency in Malayalam.
• High accuracy with Malayalam script and diacritics.
• Experience with annotation, transcription, localization, or data review.

Responsibilities

• Review assigned Malayalam pages for readability and content.
• Identify and bound every meaningful region: titles, headings, paragraphs, tables, figures, diagrams.
• Assign correct component type and reading-order index to each region.
• Record structural relationships by linking regions to figures or tables via parent IDs.
• Transcribe all visible text exactly as it appears in Malayalam script.
• Flag regions that cannot be read reliably.
• Capture page metadata including language, document type, source, and dimensions.
• Review another Malayalam expert's work end to end when assigned to reviewer role.
• Apply the annotation taxonomy consistently across pages.
• Flag unsuitable pages before annotation begins.

Skills

Malayalam transcription

Job description

Help build high-quality document training data from real Malayalam-language PDFs by mapping page structure and transcribing every text region accurately in the original script. The work focuses on difficult document types and layouts that document-understanding systems often handle poorly, including handwriting, dense multi-column pages, tables, diagrams, and mixed-script content.

Role Overview

Pages are drawn from publicly available Malayalam materials such as newspapers, textbooks, examinations, flyers, forms, manuals, menus, brochures, notices, and worksheets. Each page must be structurally annotated and fully transcribed by a human. All annotation and transcription work is performed in Malayalam, with no transliteration used for Malayalam text.

Key Responsibilities

• Review assigned PDF pages before annotation to confirm they are in Malayalam, legible, contain substantive content, and do not display personal information.
• Identify and bound every meaningful page region, including document titles, section headings, paragraphs, lists, tables, figures, diagrams, captions, formulas, questions, and answer fields.
• Assign the correct component type and reading-order index to each identified region.
• Record structural relationships by linking regions to the relevant figure or table through parent component identifiers.
• Transcribe all visible text exactly as it appears in Malayalam script, including handwritten material, and flag regions that cannot be read reliably.
• Capture page metadata including language, document type, source, page dimensions, and indicators for tables, formulas, and handwriting.
• Review another Malayalam expert''s completed work end to end when assigned to the reviewer role.
• Apply the annotation taxonomy consistently across large numbers of pages.
• Flag unsuitable pages before spending significant time annotating them.

Qualifications

Native fluency in Malayalam is required.Full command of Malayalam script and orthography, including diacritics and conjunct forms, is required.Experience working with documents through annotation, transcription, translation, localization, subtitling, proofreading, journalism, or regional-language data review.Strong attention to character-level accuracy, including correct use of Malayalam diacritics and character forms.Ability to follow a defined taxonomy consistently rather than changing annotation conventions from document to document.Comfort working with unfamiliar or complex layouts such as multi-column newspapers, examination papers, and handwritten forms.

Preferred Experience

Regional-language AI data work, including annotation, labeling, grading, or bilingual evaluation for training datasets.Transcription or localization experience such as machine-translation post-editing, subtitling, bilingual quality assurance, OCR correction, or post-editing.Document production experience, including typesetting, copy-editing, proofreading, or digitization of Malayalam-language material.Knowledge of Unicode normalization, Malayalam input methods, numeral forms, and character-form accuracy.

Quality Expectations

Capture every meaningful region of each page with accurate boundaries and component types.Set reading order to reflect how the document is actually read, including movement across columns.Match source text character for character in Malayalam script.Produce work that can pass second-expert review without correction.Prioritize accuracy while maintaining efficient handling time.

Work Terms

Remote, hourly work.PDF pages are provided directly, so annotators do not need to source documents themselves.Completed work is human-authored throughout, including component identification, component typing, reading order, and transcription.

Compensation

Hourly compensation is $10 to $11.

Eligibility

Candidates must be native Malayalam speakers with full command of Malayalam script and orthography. All annotation and transcription tasks are completed in Malayalam.