Mercor is a leading AI‑training partner that collaborates with top research labs and enterprises to build high‑quality datasets for next‑generation document understanding models. The company focuses on under‑represented scripts and languages, creating diverse, human‑verified training data that powers commercial and academic AI solutions. With a distributed workforce and a strong emphasis on quality over speed, Mercor empowers specialists to contribute to cutting‑edge research while enjoying flexible, remote work arrangements. The role of PDF Annotation & Transcription Expert – Telugu is pivotal in expanding the corpus of Telugu documents for machine learning. You will source publicly available Telugu PDFs that contain complex layouts such as multi‑column newspapers, exam sheets, and handwritten forms. Your primary responsibility is to produce a detailed structural map of each page, identifying every meaningful region—titles, headings, paragraphs, tables, figures, captions, formulas, and handwritten content. Each region must be bounded, typed, and assigned a reading‑order index, while also linking sub‑components to their parent figures or tables. Accuracy is paramount; a single diacritic error is considered a defect. After annotation, you will transcribe the text verbatim in Telugu script, flagging illegible areas. Your work will be reviewed by a second expert to ensure consistency and correctness. Key responsibilities include: 1. Locating and verifying publicly accessible Telugu PDFs across diverse domains. 2. Identifying and bounding every meaningful page region. 3. Assigning component types and reading‑order indices. 4. Linking related regions through parent identifiers. 5. Transcribing text accurately in Telugu script, including handwritten content. 6. Capturing page metadata such as language, dimensions, and special flags. 7. Reviewing peer annotations and providing constructive feedback. 8. Maintaining consistency with the established taxonomy across hundreds of pages. 9. Reporting any layout or content anomalies. 10. Contributing to the expansion of the dataset with diverse document templates. The ideal candidate is a native Telugu speaker with a strong command of the script, experience in document annotation or localization, and meticulous attention to detail. Familiarity with OCR post‑editing, Unicode normalization, and regional‑language AI datasets is a plus. Growth path: Successful annotators can progress to senior annotation leads, project managers, or data quality specialists within Mercor’s AI training ecosystem. The role offers exposure to advanced document AI research, opportunities to influence model training pipelines, and a chance to work with a global team of language experts. Why join Mercor? You’ll be at the forefront of building AI that understands complex scripts, work from anywhere, and earn competitive hourly rates. The company values quality, offers flexible schedules, and provides a supportive environment for continuous learning and skill development.