Data Labelling Engineer

Responsibilities:

1. Support data classification and taxonomy methods and standards, understand business, cooperate with data team. Support analysis, identification and marking various data samples in storage, and ensure accuracy and recall rate.

2. Support R&D team to improve algorithm recognition and machine learning ability, analyze badcase of algorithm recommended data labels, put forward suggestions and basis for annotation/labelling, and participate in the discussion of algorithm model improvement solutions

3. Deeply understand the business, investigate the data samples of each business line, and improve the classification standard of data labels and sample database

4. Improve the data annotation operation process, optimize the label sampling and review mechanism, and improve the efficiency of manual standards

5. Participate in the functional optimization design of the annotation platform

Required skills:

• Business Understanding: Deeply understands the business context and end-user scenarios behind labeling work. Able to translate business goals into annotation strategies and priorities. Demonstrates strategic interpretation of data labeling needs, rather than just process execution or task repetition.
• Guideline Development: Proven experience in designing, evolving, and owning annotation guidelines from scratch. Capable of abstracting rules from ambiguous inputs, formalizing standards, and continuously refining them based on data insights and product feedback. Acts as a thought leader in standard-setting, not just a rule follower.
• Quality Improvement: Possesses a systematic, data-driven methodology for quality management—including defining bad-case taxonomies, attributing root causes, designing sampling and re-review mechanisms, and establishing closed-loop feedback systems. Proactively owns quality outcomes and drives continuous improvement across the team.
• Cross-Functional Collaboration: Demonstrates strong ownership in identifying gaps, initiating discussions, aligning stakeholders (engineering, product, senior analysts), and driving solutions from proposal to execution. Serves as the problem owner, not just an execution node in the workflow.
• Structured Thinking & Communication: Articulates ideas clearly and concisely, with a framework-driven approach. Able to abstract complex problems into well-defined analytical frameworks, decision criteria, and outcome-driven summaries. Communicates with clarity, logic, and strategic intent—both verbally and in writing.