> Markdown version of [/jobs/ext/4682283-data-analyst-japanese](https://www.wearedevelopers.com/jobs/ext/4682283-data-analyst-japanese). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md). --- # Data Analyst (Japanese) - **Company:** Ririo.Com, Inc. - **Location:** United States - **Experience:** Expert - **Contract:** Permanent contract - **Skills:** Artificial Intelligence, Data Files, Information Technology, Data Analytics, Evaluation of Large Language Models - **Published:** November 12, 2026 - **Apply:** https://www.indeed.com/viewjob?jk=7e38be559469351e ## About the Role Native-level fluency with deep, current cultural immersion and living experience in the target market (Japanese - Japan). * Demonstrated experience developing, adapting, or refining content policies and safety guidelines in the target language/market. * Experience applying or working with harm and risk taxonomies, classification frameworks, or core safety policy principles. * Proven ability to analyze nuanced language: identifying market-specific slang, regional idioms, double entendres, veiled speech, and cultural sensitivities. * Professional background in Linguistics, Translation/Localization, Trust & Safety, Content Moderation, or AI Policy Evaluation. * Experience evaluating sensitive, policy-violating, or Responsible AI content. * Experience working with LLM evaluation, prompt assessment, or data labeling workflows. * Bachelor's degree in Data Science, Computer Science, Data Analytics or a related field ## Description * Guideline & Policy Development: Partner with other teams to draft, adapt, and refine market-specific safety guidelines, and policy definitions for the target locale. * Linguistic & Cultural Contextualization: Translate abstract safety policies into actionable local standards, providing clear definitions and localized examples of edge cases, regional slang, euphemisms, and indirect harmful intent. * Exemplar Dataset Creation: Develop and validate high-quality, culturally authentic ground-truth (exemplar) datasets to calibrate autograders and guide third-party annotation teams. * Prompt & Evaluation Adaptation: Adapt English safety prompts and test sets into culturally natural, dialectally accurate language rather than literal translations. * Model Failure Analysis: Investigate safety-related model failures, diagnosing whether errors stem from regional dialectal quirks, coded language, or cultural misunderstandings, and update guidelines accordingly. * Regional Safety Advisory: Act as the subject matter expert on local sensitive topics, emerging internet culture, social context, and regulatory or cultural taboos for the target region.