General Knowledge
Create or evaluate accurate, well-explained content across everyday, cultural, educational, and public-knowledge topics.
Indic Language & Domain Expert — AI Training and Evaluation
Paid remote contract work for native Indic-language professionals with proven subject expertise. Create, review, and evaluate AI training tasks in the language and domain combinations you genuinely know well.
Apply for this role →What you'll do
- Create, solve, review, or evaluate AI-training tasks and model responses in your approved language and domain combinations.
- Identify linguistic, factual, reasoning, mathematical, coding, instruction-following, reading-comprehension, or tool-use errors relevant to your selected track.
- Write precise corrections, reference answers, rationales, or scoring notes that another qualified reviewer can audit.
- Follow project rubrics and calibration feedback consistently, protect confidential material, and flag ambiguous or unanswerable tasks.
- Accept work only for the language-category combinations for which you have been assessed and approved.
What we look for
- Native fluency in at least one listed language, including strong professional reading and writing ability in its standard script.
- B1-B2 or stronger English proficiency, sufficient to understand detailed instructions, source material, and quality feedback.
- Demonstrated expertise in at least one listed domain through professional work, research, teaching, formal education, publications, projects, or a credible portfolio.
- Ability to explain judgements clearly, work independently in browser-based tools, and maintain accuracy across repeated tasks.
- Willingness to complete separate language and domain assessments before being matched to paid project work.
Eligible languages
Bengali
Gujarati
Hindi
Kannada
Malayalam
Marathi
Nepali
Odia
Punjabi
Tamil
Telugu
Urdu
Choose your expertise track
Commonsense
Judge real-world plausibility, implicit assumptions, cause and effect, and the practical logic behind an answer.
Math
Solve and verify mathematical problems, calculations, reasoning steps, and final answers at the level supported by your background.
Code
Write, review, debug, or explain code and reason precisely about algorithms, software behaviour, tests, and edge cases.
Chat & Instruction
Evaluate natural dialogue, instruction following, clarity, helpfulness, tone, and whether a response fully satisfies the user's request.
Agentic, Code and Tool Call
Review multi-step agent behaviour, code-assisted workflows, tool selection, tool arguments, execution traces, and task completion.
Reading
Interpret passages in the target language, answer questions from the text, and distinguish supported conclusions from unsupported claims.
Long Context (English Mirror)
Work from long English source material and create or evaluate target-language outputs while preserving meaning, constraints, and important details.
Before you apply
Apply only for combinations you can support professionally. Conversational familiarity, ancestry, residence, or listing a language on a resume is not enough. Each language-domain combination is screened separately, and you may be asked to provide a short native-language sample and category-specific evidence or assessment.