Review and correct AI-processed documents to help improve the next generation of AI-powered document understanding tools. You’ll work with text, tables, layouts, and other document elements, making precise corrections so the resulting training data is accurate and consistent.
About this project
Project LIPI focuses on improving how AI systems understand complex documents. You’ll review AI-tagged and transcribed content, correct errors, and ensure document structures such as text blocks, tables, titles, images, and formulas are accurately represented.
This is detail-focused work where accuracy matters more than speed.
What you’ll do
- Review AI-tagged document layouts, including text blocks, tables, titles, and images, and correct inaccurate annotations.
- Review transcribed text and correct errors, including language- and script-specific details.
- Structure tables and formulas accurately for use as AI training data.
- Work with LaTeX when required to represent formulas and structured content. Training will be provided.
- Follow detailed written guidelines to ensure consistent, high-quality annotations.
- Flag unclear or ambiguous content instead of guessing.
Requirements
- Native-level fluency in one of the eligible languages and residence in the corresponding target country.
- Strong reading and writing skills in your selected language.
- Ability to pass a language proficiency assessment before onboarding.
- Excellent attention to detail and a strong focus on accuracy.
- Ability to follow detailed written instructions consistently.
- A computer with a stable internet connection.
- Previous experience with annotation, translation, proofreading, or transcription is a plus.
- Experience working with legal or government documents, tables, or other structured data is a plus.
Eligible languages
This project currently supports:
- Assamese
- Bengali
- Gujarati
- Hindi
- Kannada
- Malayalam
- Marathi
- Oriya
- Punjabi
- Tamil
- Telugu
- Urdu
Applicants must be native speakers of their selected language and currently located in India.
Project details
- Location: Remote
- Project type: Annotation
- Duration: Work may be available throughout the project period, depending on project needs and task availability.
- Assessment: Candidates must successfully complete a language proficiency assessment before onboarding.
