Hindi QA Lead

Role overview

This remote, hourly contractor position is for a Hindi Quality Assurance Lead who will own quality control and consistency across Hindi AI training work. The role centers on reviewing AI-generated Hindi output as well as trainer and QA deliverables, measuring them against project rubrics, and giving clear written feedback that helps raise the overall standard of the work.

You will assess content for correctness, fluency, grammar, spelling, tone, cultural fit, meaning accuracy, instruction compliance, formatting, and alignment with project requirements. In addition, you will look for repeated quality problems, share updates with trainers and QAs, help with onboarding, keep process documentation current, and support the reactivation of contributors who are not consistently active.

This is with a fast-scaling AI data services company that provides training data to major AI companies and foundation model teams. Strong Hindi quality leadership in this role will help improve Hindi model outputs by making sure the data is natural, accurate, culturally appropriate, well documented, and consistent with client expectations.

There is no active project attached to this role right now, but qualified candidates may be contacted first when relevant work becomes available. The role also gives access to future opportunities through the expert network.

Selection process

Candidates should expect an AI interview, a domain-specific task, and a final interview with a recruiter.

Responsibilities

• Review Hindi items for quality, raise issues, and share feedback directly with contributors through DMs.
• Flag repeat or serious problems so they can be addressed quickly.
• Keep trainers and QAs informed in Discord about guideline changes, workflow updates, project shifts, and quality expectations.
• Answer questions on Hindi grammar, Devanagari script, register, translation accuracy, Hinglish or code-switching, terminology, and cultural context.
• Contact inactive contributors, encourage them to resume work, monitor follow-ups, and report availability concerns.
• Build and maintain Hindi style guides, trackers, FAQs, examples, honeypots, onboarding resources, and related documentation.
• Lead onboarding and training calls to explain workflows, rubrics, quality standards, and Hindi-specific requirements.
• Apply Hindi language rules consistently across the project.
• Detect recurring quality gaps and help improve QA processes.

Requirements

• A bachelor’s or master’s degree in Hindi, Linguistics, Translation, Communications, Journalism, English, Education, Quality Assurance, or a related field.
• Native or near-native Hindi ability with excellent reading and writing skills.
• Strong English skills for understanding guidelines, collaborating with teams, and writing feedback clearly.
• At least 3 years of professional experience in Hindi writing, editing, translation, localization, content QA, AI training, education, annotation, or a similar language-review workflow.
• Solid command of Hindi grammar, Devanagari script, spelling, punctuation, tone, register, formal and conversational usage, and India-specific cultural context.
• Ability to judge Hindi content using detailed rubrics and spot issues such as mistranslation, overly literal wording, unnatural phrasing, incorrect gender or verb agreement, fabricated claims, ambiguity, or inconsistent terminology.
• Prior experience leading or supporting remote teams of trainers, annotators, reviewers, editors, or QAs is preferred.
• Comfort using Discord, Google Sheets, Google Docs, trackers, dashboards, and project management tools.
• Strong organization and attention to detail, with the ability to manage style guides, FAQs, trackers, onboarding materials, honeypots, and documentation.
• Experience in AI training, data annotation, LLM evaluation, prompt/response review, or rubric-based LLM QA is a strong advantage.

Additional information

This is an hourly remote contract position. The work is focused on Hindi AI training projects and quality leadership for language output.