About the role
LYANAI supports real dating and social conversations. High-quality data is essential to understanding context, assessing appropriate expression and improving replies.
We are looking for someone who reads conversations carefully, distinguishes tone, intent and information boundaries, and works with our algorithm and product teams to improve annotation consistency and data quality.
Responsibilities
- Organise and annotate dialogue data following guidelines, identifying context, intent, emotions and reply relevance.
- Evaluate AI replies and compare preferences, considering naturalness, appropriateness, factual support and conversational coherence.
- Clean and deduplicate data, identify unusual samples and conduct spot checks, recording issues and the reasoning behind your decisions.
- Collect ambiguous or disputed examples and work with the team to improve guidelines, edge cases and example libraries.
- Maintain evaluation datasets with the algorithm and product teams, track annotation quality and report issues from real scenarios.
- Follow data authorisation, de-identification and access rules. Do not collect, share or reuse personal conversations or sensitive information without permission.
Requirements
- Strong Chinese reading and writing skills, with the ability to distinguish explicit information, reasonable inference and speculation in context.
- Attention to detail and patience, following shared standards and maintaining consistent judgement in repetitive tasks.
- The ability to clearly record issues and reasoning, and communicate promptly about ambiguous or disputed examples.
- Familiarity with document and spreadsheet tools, and a willingness to learn annotation platforms and basic data processing.
- An interest in AI products, language and real human communication, with respect for different expression styles and personal boundaries.
Nice to have
- Experience in text or dialogue annotation, content moderation or model reply evaluation.
- Knowledge of preference annotation, annotation consistency or data quality sampling.
- Knowledge of linguistics or psychology, or the ability to use scripts for basic data preparation.
How to apply
Include “Application: Data Annotation Engineer” in your email subject, attach your CV and briefly describe your relevant experience.
You are welcome to include public examples of annotation guidelines, quality reviews or text analysis. Do not submit private information or work data you are not authorised to share.