{"schemaVersion":"jobsearcher.job.v1","id":"d62060ed4b294a15afaabaef","url":"https://jobsearcher.com/jobs/d62060ed4b294a15afaabaef","canonicalUrl":"https://jobsearcher.com/jobs/d62060ed4b294a15afaabaef","title":"Post-Training Data Preference Data Reviewer","description":"Post-Training Data Preference Data Reviewer is a remote evaluation track for reviewing post training data preference data evaluation prompts and responses against AuraOne's quality rubric. Reviewers compare paired outputs, label edge cases, and write the kind of structured feedback the modeling team can use to retrain.\nWhy this role matters\nAI data reviewers help turn post training data preference data evaluation outputs into auditable labels, rationales, and regression cases for AuraOne Human Data.\nResponsibilities\nEvaluate post training data preference data evaluation model outputs against a versioned rubric and assign severity tags for Post-Training Data Preference Data Reviewer assignments.\nCompare paired responses and pick the stronger answer with a written rationale.\nLabel hallucinations, instruction-following failures, and unsafe content with structured tags.\nCapture ambiguous prompts and route them back to the program team for rubric updates.\nMaintain reviewer-quality scores by calibrating against gold-standard examples each week.\nDocument recurring failure modes so the modeling team can target them in the next training run.\nQualifications\nPrior evaluation, annotation, or human-rater experience on post training data preference data evaluation or adjacent content for Post-Training Data Preference Data Reviewer work.\nComfort applying multi-page rubrics consistently across long batches.\nClear written reasoning that names the issue and the rubric clause being applied.\nStrong attention to detail and the ability to flag when a prompt itself is the problem.\nReliable async availability for at least 10 hours per week.\nExample tasks\nCompare two post training data preference data evaluation model responses to the same prompt and pick the stronger one with rationale.\nTag an unsafe response with the correct policy category and severity.\nAudit a 50-row batch for rubric consistency and report drift to the program lead.\nPropose a rubric clarification after spotting a recurring failure mode.\nNice to have\nBackground in linguistics, content moderation, or trust & safety review.\nExperience with inter-rater agreement metrics and calibration cycles.\nDomain expertise that lets you spot subject-matter errors automated checks miss.\nSkills\nModel output evaluation\nRubric-based annotation\nSeverity tagging\nInter-rater calibration\nPost Training Data Preference Data evaluation\nPreference ranking\nRLHF\nRater calibration\nPost\nTraining\nData\nWork model\nRemote — US-eligible. Remote · Independent specialist contractor. Employment type: CONTRACTOR. Applicants must be authorized to work from US.\nCompensation\nHourly rate confirmed after the interview process.\nApplication process\nApply through AuraOne's specialist intake for role-specific routing and review. Final project scope, schedule, and contractor terms are confirmed before placement.","company":"Auraone Human Data","rawCompany":"auraone human data","city":"Remote","state":"OR","isRemote":false,"isActive":false,"createdAt":"2026-08-04T23:26:58.040Z","occupations":[{"code":"43-9081.00","title":"Proofreaders and Copy Markers","slug":"proofreaders-and-copy-markers"},{"code":"13-1151.00","title":"Training and Development Specialists","slug":"training-and-development-specialists"},{"code":"15-2051.00","title":"Data Scientists","slug":"data-scientists"}],"industries":[{"code":"541690","title":"Other Scientific and Technical Consulting Services","slug":"other-scientific-and-technical-consulting-services"},{"code":"541618","title":"Other Management Consulting Services","slug":"other-management-consulting-services"},{"code":"541511","title":"Custom Computer Programming Services","slug":"custom-computer-programming-services"}],"jobPosting":{"@context":"https://schema.org","@type":"JobPosting","title":"Post-Training Data Preference Data Reviewer","description":"Post-Training Data Preference Data Reviewer is a remote evaluation track for reviewing post training data preference data evaluation prompts and responses against AuraOne's quality rubric. Reviewers compare paired outputs, label edge cases, and write the kind of structured feedback the modeling team can use to retrain.\nWhy this role matters\nAI data reviewers help turn post training data preference data evaluation outputs into auditable labels, rationales, and regression cases for AuraOne Human Data.\nResponsibilities\nEvaluate post training data preference data evaluation model outputs against a versioned rubric and assign severity tags for Post-Training Data Preference Data Reviewer assignments.\nCompare paired responses and pick the stronger answer with a written rationale.\nLabel hallucinations, instruction-following failures, and unsafe content with structured tags.\nCapture ambiguous prompts and route them back to the program team for rubric updates.\nMaintain reviewer-quality scores by calibrating against gold-standard examples each week.\nDocument recurring failure modes so the modeling team can target them in the next training run.\nQualifications\nPrior evaluation, annotation, or human-rater experience on post training data preference data evaluation or adjacent content for Post-Training Data Preference Data Reviewer work.\nComfort applying multi-page rubrics consistently across long batches.\nClear written reasoning that names the issue and the rubric clause being applied.\nStrong attention to detail and the ability to flag when a prompt itself is the problem.\nReliable async availability for at least 10 hours per week.\nExample tasks\nCompare two post training data preference data evaluation model responses to the same prompt and pick the stronger one with rationale.\nTag an unsafe response with the correct policy category and severity.\nAudit a 50-row batch for rubric consistency and report drift to the program lead.\nPropose a rubric clarification after spotting a recurring failure mode.\nNice to have\nBackground in linguistics, content moderation, or trust & safety review.\nExperience with inter-rater agreement metrics and calibration cycles.\nDomain expertise that lets you spot subject-matter errors automated checks miss.\nSkills\nModel output evaluation\nRubric-based annotation\nSeverity tagging\nInter-rater calibration\nPost Training Data Preference Data evaluation\nPreference ranking\nRLHF\nRater calibration\nPost\nTraining\nData\nWork model\nRemote — US-eligible. Remote · Independent specialist contractor. Employment type: CONTRACTOR. Applicants must be authorized to work from US.\nCompensation\nHourly rate confirmed after the interview process.\nApplication process\nApply through AuraOne's specialist intake for role-specific routing and review. Final project scope, schedule, and contractor terms are confirmed before placement.","datePosted":"2026-08-04T23:26:58.040Z","dateModified":"2026-08-04T23:26:58.040Z","hiringOrganization":{"@type":"Organization","name":"Auraone Human Data","sameAs":"https://jobsearcher.com"},"jobLocation":{"@type":"Place","address":{"@type":"PostalAddress","addressLocality":"Remote","addressRegion":"OR","addressCountry":"US"}},"identifier":{"@type":"PropertyValue","name":"JobSearcher","value":"d62060ed4b294a15afaabaef"},"url":"https://jobsearcher.com/jobs/d62060ed4b294a15afaabaef"}}