Job Overview
Location
South Korea
Job Type
Full-time
Category
Software Engineering
Date Posted
March 26, 2026
Full Job Description
đź“‹ Description
- • This role is a critical contributor to a high-impact AI safety initiative, where bilingual evaluators in South Korea help ensure AI-generated responses are culturally appropriate, safe, and reliable for Korean-speaking users. By applying a structured safety rubric to prompt-image pairs, you directly support the development of trustworthy AI systems that respect local norms and mitigate harm—making your work essential to ethical AI deployment in one of Asia’s most technologically advanced markets.
- • You will evaluate AI-generated responses using a detailed safety rubric, completing two independent assessments per prompt-image pair to ensure consistency and reduce bias. Your day-to-day work involves analyzing nuanced cultural context, identifying potential risks in model outputs (such as harmful, offensive, or misleading content), and providing clear, well-structured rationales in English for each evaluation. You will participate in calibration sessions to align with team standards, support arbitration when discrepancies arise between evaluators, and maintain strict quality and throughput targets—aiming for 20–40 hours per week over a 2–3 week engagement with flexible scheduling options (either 4 hours/day Mon–Fri or 2 hours/day Mon–Fri plus 10 weekend hours).
- • You will join Welo Data, a specialized division of Welocalize, Inc., a global leader in AI training data with over 25 years of experience serving innovative technology companies. Welo Data leverages a curated network of over 500,000 domain experts worldwide to deliver high-quality, ethically sourced, and scalable datasets for LLMs, search, ads, and multimodal AI systems. As part of this project, you’ll collaborate with a remote, multicultural team of AI safety specialists focused on ensuring models behave responsibly across linguistic and cultural boundaries.
- • This role offers a unique opportunity to deepen your expertise in AI safety, content evaluation, and cross-cultural assessment—skills that are increasingly in demand as AI systems scale globally. You will gain hands-on experience applying safety frameworks to real-world model outputs, refine your analytical and written communication skills in English, and contribute to meaningful projects that shape how AI interacts with diverse communities. Successful completion may lead to future opportunities on similar multilingual AI evaluation projects, particularly if you demonstrate reliability, precision, and cultural insight.
🎯 Requirements
- • Fluency in Korean and English, with native or near-native proficiency in Korean
- • Deep cultural understanding of South Korean social norms, sensitivities, and contextual nuances relevant to AI safety evaluation
- • Prior experience in safety evaluation, content moderation, policy review, or rubric-based assessment (preferred)
- • Strong written English skills to produce clear, concise, and well-structured rationales for evaluations
- • Ability to apply detailed guidelines consistently and maintain high attention to nuance and detail
- • Reliable availability to commit 20–40 hours per week during the 2–3 week project window
🏖️ Benefits
- • Competitive hourly rate of 22 USD for remote freelance work
- • Fully remote engagement—work from anywhere in South Korea with flexible scheduling options
- • Opportunity to contribute to ethically significant AI safety projects with real-world impact on model reliability
- • Gain specialized experience in AI evaluation, cultural assessment, and LLM safety—valuable for careers in AI ethics, data annotation, or tech policy
- • Collaborate with a global team of AI experts at Welocalize, a trusted partner to leading technology companies with 25+ years of industry experience
Skills & Technologies
About Welocalize, Inc.
Welocalize, Inc. is a global language services and technology company that provides translation, localization, and content transformation solutions. Founded in 1997 and headquartered in Frederick, Maryland, the company helps enterprises adapt products, marketing, support, and digital experiences for more than 250 languages. Services include technical translation, multilingual SEO, machine-translation post-editing, data collection, linguistic testing, and AI training data. Welocalize operates a worldwide network of linguists and engineers, supported by proprietary automation and quality-management platforms. Clients span technology, life sciences, finance, manufacturing, and consumer sectors, enabling international growth through culturally relevant, compliant, and scalable multilingual content.
Subscribe to the weekly newsletter for similar remote roles and curated hiring updates.
Newsletter
Weekly remote jobs and featured talent.
No spam. Only curated remote roles and product updates. You can unsubscribe anytime.
Similar Opportunities
SHI International Corp.
2 months ago
Aquia Inc.
7 months ago
19 days ago