Native-Speaker OCR Annotator/Reviewer

Role & responsibilities

• Verify that bounding boxes accurately cover each meaningful block on the page (text, titles, tables, pictures, formulas, etc.).
• Confirm the correct class label is applied to each block, using the approved taxonomy (Text, Title, Section-header, List-item, TOC, Bibliography, Footnote, Page-header, Page-footer, Picture, Formula, Table, Caption).
• Check and correct reading order so blocks are sequenced the way a person would naturally read the page.
• Review and correct transcribed text for accuracy, ensuring fidelity to the source including superscripts/subscripts, bracket styles, numbering, and redacted/omitted text markers.
• Verify tables are correctly structured in LaTeX (rows, columns, merged cells) and formulas are correctly rendered in LaTeX.
• Ensure transcription follows Unicode normalization standards and pays close attention to script-specific details (conjuncts, ligatures, matra placement).
• Flag ambiguous or undocumented cases (e.g., unclear symbols, damaged scans, edge cases not covered by guidelines) for spec-owner review rather than guessing.
• Maintain a review-pass rate of at least 95%, and complete a second-pass/native-reader verification before final delivery.

Preferred candidate profile

• Native fluency in the target language (reading and writing), with strong comprehension of the script's grammar and typographic conventions.
• Strong attention to detail comfortable with repetitive, precision-focused review work.
• Basic comfort with digital annotation tools (training will be provided on the specific platform).
• Familiarity with LaTeX is a plus, but not mandatory (training provided for table/formula annotation).
• Prior experience in data annotation, proofreading, translation, or transcription work is preferred but not required.

Ability to follow detailed written guidelines precisely and consistently

• Experience working with scanned/legal/government documents (gazettes, notifications, official records).
• Familiarity with OCR, NLP, or AI training data concepts.

Experience reviewing tabular or structured data formats