Telugu Document Annotation Pro

Mercor; Zürich; Vor Ort; CHF 50.000 - 73.000; Vollzeit

Zusammenfassung

Mercor is seeking meticulous Telugu language annotators to build a diverse dataset from real Telugu documents, including newspapers, forms and exams. You will bind and label each region, determine reading order, and ensure pixel-accurate transcription in Telugu script.

Collaboration with a second expert ensures quality at every step. You are a native Telugu speaker with exacting accuracy, comfortable with complex layouts, and experienced in document-related work such as annotation,

Qualifikationen

• Must be a native Telugu speaker with full script knowledge.
• Experience with document annotation, transcription, localization or QA is desirable.
• Character-level accuracy and correct diacritics are essential.
• Apply a consistent taxonomy across hundreds of pages.
• Comfort with multi-column layouts, handwritten forms, and varied documents.

Aufgaben

• Source publicly accessible Telugu PDFs with multimodal elements and record the source.
• Annotate structure: bound regions, assign type and reading-order index.
• Link regions to their parent figure or table.
• Transcribe faithfully in Telugu script, flag unreadable text.
• Capture page metadata: language, type, source, dimensions, handwriting/diagrams flags.
• Participate in peer review with a second Telugu expert.

Kenntnisse

Native Telugu speaker

Script fluency

Attention to detail

Document annotation experience

Systematic approach

Reading-order discipline

Tools

Unicode normalization

Jobbeschreibung

Mercor is seeking meticulous Telugu language annotators to build a diverse dataset from real Telugu documents, including newspapers, forms and exams. You will bind and label each region, determine reading order, and ensure pixel-accurate transcription in Telugu script.

Collaboration with a second expert ensures quality at every step. You are a native Telugu speaker with exacting accuracy, comfortable with complex layouts, and experienced in document-related work such as annotation,