你是否热爱语言、写作和质量评估？ 我们正在启动一个全新的语言评估项目，现诚邀具备普通话能力的专业人士加入团队。 在这个项目中，你将通过结构化的评估与写作任务，帮助训练 AI 系统更好地理解、生成和优化语言内容。工作可全程远程完成，并支持灵活安排时间。 你将负责 评估 AI 生成的回答，检查其准确性、语法以及文化语境的适切性。 识别内容中的问题，并提供经过优化的高质量改写版本。 使用普通话撰写自然的提示词（Prompt）和回答，帮助提升对话数据集的质量。 与全球团队协作，共同推动 AI 语言模型的持续改进。 项目详情 职位名称： Project Lyra…
Marathi Expert with Mathematical Background
India, Mauritius, Pakistan
Up to $20/hr
August 14, 2026
Job description
In this hourly, remote contractor role, you will work as a Marathi Language Specialist (SME) to review AI-generated Marathi content and/or create high-quality Marathi reference content, evaluating language quality and step-by-step reasoning while providing precise written feedback. You will assess outputs for accuracy, fluency, grammar, spelling, tone, cultural appropriateness, and adherence to the prompt; identify errors in meaning, mistranslation, register (formal vs informal), ambiguity, and logical consistency; fact-check claims when needed; write high-quality rewrites and model answers that demonstrate best practices in Marathi writing; and rate and compare multiple AI responses based on correctness and overall quality. This role is with SME Careers, a fast-growing AI Data Services company and subsidiary of SuperAnnotate, delivering training data for many of the world’s largest AI companies and foundation-model labs. Your Marathi expertise directly helps improve the world’s premier AI models by making their Marathi language output more natural, reliable, culturally accurate, and clearly written.
Responsibilities
- Develop AI Training Content: Create prompts and gold-standard Marathi answers across multiple domains (general knowledge, customer support, education, conversational writing).
- Optimize AI Performance: Evaluate and rank AI outputs to improve fluency, correctness, tone match, and cultural appropriateness.
- Ensure Model Integrity: Detect unsafe or misleading content, factual hallucinations, bias, and policy-risk language; validate reliability across use cases.
Requirements
- Bachelor’s degree (or higher) in Mathematics
- Native or near-native Marathi proficiency (reading and writing required).
- Strong professional proficiency in English (minimum C1) to follow guidelines and provide detailed feedback in English.
- 3+ years of professional experience in writing, editing, localization, translation, content QA, journalism, education, or related language-focused work.
- Deep understanding of Marathi grammar, spelling conventions, punctuation, and style—able to explain why something is wrong and how to fix it.
- Comfortable handling multiple registers (formal/informal), dialect variation (e.g., Pune/Mumbai/standard usage vs regional influences), and audience adaptation (India vs diaspora).
- High attention to detail when spotting subtle meaning shifts, awkward literal translations, hallucinated claims, or culturally mismatched phrasing.
- Reliable, self-directed, and able to deliver consistent quality in an hourly, remote contractor workflow across time zones.
- Prior experience with AI data training/annotation, translation QA, editorial QA, or rubric-based evaluation is strongly preferred.
Compensation by region
Pay varies by where you're based. Each region below is open to applicants in that country only.
Apply now
Opens sme.careers in a new tab. If they ask how you found the role, mention RemoWork.
你是否热爱语言、写作和质量评估？ 我们正在启动一个全新的语言评估项目，现诚邀具备普通话能力的专业人士加入团队。 在这个项目中，你将通过结构化的评估与写作任务，帮助训练 AI 系统更好地理解、生成和优化语言内容。工作可全程远程完成，并支持灵活安排时间。 你将负责 评估 AI 生成的回答，检查其准确性、语法以及文化语境的适切性。 识别内容中的问题，并提供经过优化的高质量改写版本。 使用普通话撰写自然的提示词（Prompt）和回答，帮助提升对话数据集的质量。 与全球团队协作，共同推动 AI 语言模型的持续改进。 项目详情 职位名称： Project Lyra…
Productive Playhouse is building a talent pool of Spanish (LatAm) speakers for an upcoming project testing and evaluating leading AI chatbots. The objective is to enhance response quality through direct user interaction with the various AI models.
Productive Playhouse is building a talent pool of Egyptian Arabic speakers for an upcoming project testing and evaluating leading AI chatbots. The objective is to enhance response quality through direct user interaction with the various AI models.