Remote Cross-Modal Evaluation Specialist

AuraOne; United States; On-site; USD 34,440 - 55,104; Part time

Job summary

AuraOne is seeking a remote Cross-Modal Retrieval Model Evaluator to review prompts and model outputs against a quality rubric. You will compare paired responses, label edge cases, and provide structured feedback for the modeling team to retrain. This contractor role requires careful labeling, documenting recurring failure modes, and maintaining calibration against gold standards each week. Applicants should be authorized to work from the United States and comfortable with async collaboration.

Qualifications

• Experience evaluating cross-modal model outputs against rubrics.
• Ability to name issues and map them to rubric clauses.
• Strong attention to detail and written reasoning.

Responsibilities

• Evaluate cross-modal retrieval model evaluation outputs against a versioned rubric and assign severity tags for assignments.
• Compare paired responses and select 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.
• Document recurring failure modes so the modeling team can target them in the next training run.

Skills

Model output evaluation

Rubric-based annotation

Severity tagging

Inter-rater calibration

Cross Modal Retrieval Model evaluation

Multimodal evaluation

Cross-modal reasoning

Grounding review

Cross

Modal

Retrieval

Job description

AuraOne is seeking a remote Cross-Modal Retrieval Model Evaluator to review prompts and model outputs against a quality rubric. You will compare paired responses, label edge cases, and provide structured feedback for the modeling team to retrain. This contractor role requires careful labeling, documenting recurring failure modes, and maintaining calibration against gold standards each week. Applicants should be authorized to work from the United States and comfortable with async collaboration.