Remote Cantonese Evaluator - Data & Rubric Reviewer

AI Trainer Jobs; United States; Remote; USD 34,000 - 62,000; Part time

Benefits offered by this job

Remote work

Hourly contract

Job summary

AI Trainer Jobs is seeking a Cantonese Language Evaluator for a remote contractor role. You will review cantonese generalist evaluation prompts and responses, apply AuraOne's quality rubric, and label edge cases with structured feedback to retrain the model.

You will compare outputs, write a rationale, label hallucinations and safety issues, and calibrate scores with gold-standard examples. Reliable async availability for at least 10 hours per week and eligibility from the US are required;

Qualifications

• Prior evaluation, annotation, or human-rater experience on Cantonese generalist evaluation.
• Ability to apply multi-page rubrics consistently across long batches.
• Clear written reasoning naming the issue and rubric clause being applied.
• Reliable async availability for at least 10 hours per week.
• Experience with inter-rater agreement metrics and calibration cycles.

Responsibilities

• Evaluate cantonese generalist evaluation model outputs against a versioned rubric and assign severity tags for Cantonese Language Evaluator assignments.
• Compare paired responses and pick the stronger answer with a written rationale.
• Label hallucinations, instruction-following failures, and unsafe content with structured tags.
• Capture ambiguous prompts and route them back to the program team for rubric updates.
• Maintain reviewer-quality scores by calibrating against gold-standard examples each week.

Skills

Attention to detail

Written reasoning

Async availability

Inter-rater calibration

Cantonese evaluation

Linguistics background

Education

Linguistics background

Job description

AI Trainer Jobs is seeking a Cantonese Language Evaluator for a remote contractor role. You will review cantonese generalist evaluation prompts and responses, apply AuraOne's quality rubric, and label edge cases with structured feedback to retrain the model.

You will compare outputs, write a rationale, label hallucinations and safety issues, and calibrate scores with gold-standard examples. Reliable async availability for at least 10 hours per week and eligibility from the US are required;