LILT Cantonese AI Content Experts

📋 Description

• • Shape the future of multilingual AI: As a Cantonese AI Content Expert you will evaluate, annotate, and refine the very data that powers LILT’s next-generation neural machine translation and large-language-model stack. Every prompt, subtitle, and transcript you touch directly improves the accuracy, cultural nuance, and brand voice of translations delivered to Fortune 500 customers and government agencies.
• • Prompt evaluation & refinement: You will review English-to-Cantonese and Cantonese-to-English prompts generated by our agentic AI workflows, scoring them for fluency, terminology consistency, and contextual correctness. When a prompt misses the mark, you will rewrite or re-tag it so the model learns from precise, high-quality feedback.
• • Multimedia comprehension: Beyond text, you will watch, listen to, and transcribe short-form videos, webinars, and audio clips in Cantonese, verifying that speech-to-text outputs, captions, and on-screen text are faithful to the original meaning and appropriate for Traditional Chinese-speaking audiences.
• • Error categorization & taxonomy: Using LILT’s proprietary annotation interface, you will classify errors at granular levels—lexical, syntactic, cultural, or brand-specific—enabling our research team to target model improvements and measure quality uplift sprint-over-sprint.
• • Guideline stewardship: You will internalize evolving project guidelines and style guides, then apply them with unwavering consistency across thousands of segments. When edge cases arise, you will flag ambiguities and propose guideline clarifications that scale across the global linguist network.
• • Cross-functional collaboration: Weekly virtual stand-ups with ML engineers, product managers, and fellow linguists ensure your on-the-ground insights feed directly into roadmap decisions. You will occasionally join “linguist roundtables” to share best practices and influence tooling enhancements.
• • Continuous learning loop: LILT provides free access to its state-of-the-art NMT platform and ongoing training modules. You will experiment with new features in sandbox environments, then document what works so the entire community of freelancers benefits from your discoveries.
• • Impact metrics you will own: target throughput of 1,500 segments or 60 minutes of multimedia per week, inter-annotator agreement ≥ 90 %, and defect escape rate < 1 %. Hitting these benchmarks unlocks fast-track invitations to higher-complexity projects and long-term engagements.
• • Flexible, remote lifestyle: Work from anywhere with a stable internet connection and a quiet space for audio review. Choose the hours that fit your life—early-bird or night-owl—while still meeting weekly deadlines and quality gates.
• • Career trajectory: High-performing experts are invited to become “Gold Reviewers” who mentor newcomers, pilot beta features, and earn premium rates. Alumni have transitioned into full-time roles in data science, localization program management, and AI product design at LILT and partner organizations.