Remote AI Evaluation Auditor for Gmail & Calendar

AuraOne, Inc.; United States; Remote; USD 34,000 - 55,000; Full time

Job summary

AuraOne, Inc. seeks a remote Gmail & Calendar AI Assistant Evaluator to review prompts and responses against our quality rubric.

You will compare outputs, tag edge cases, and write structured feedback for retraining the modeling team. As an independent contractor, you will work asynchronously, at least 10 hours per week, calibrating against gold-standard examples and documenting recurring failure modes to inform future training runs.

Qualifications

• Prior evaluation, annotation, or human-rater experience on Gmail & Google Calendar AI Assistant evaluation or adjacent content.
• Ability to apply multi-page rubrics consistently.
• Clear written reasoning naming the issue and rubric clause being applied.
• Strong attention to detail and the ability to flag when a prompt itself is the problem.
• Reliable async availability for at least 10 hours per week.

Responsibilities

• Evaluate Gmail & Google Calendar AI Assistant outputs against a versioned rubric and assign severity tags.
• Compare paired responses and provide 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.
• Document recurring failure modes so the modeling team can target them in the next training run.

Skills

Evaluation experience

Rubric-based annotation

Attention to detail

Written rationale

Async availability

Gmail/Google Calendar knowledge

Job description

AuraOne, Inc. seeks a remote Gmail & Calendar AI Assistant Evaluator to review prompts and responses against our quality rubric.

You will compare outputs, tag edge cases, and write structured feedback for retraining the modeling team. As an independent contractor, you will work asynchronously, at least 10 hours per week, calibrating against gold-standard examples and documenting recurring failure modes to inform future training runs.