Search & AI Evaluation Team
Budget: $25 – $50 USD
My goal is to boost overall search accuracy across web, conversational, and voice-based platforms, and I need a small team that can run continuous quality checks on three fronts:
• First, you will rate the relevance of live web search results against real user queries, flagging mismatches and edge cases.
• Second, you will review AI-generated snippets, answers, and summaries, highlighting factual errors, bias, or tone problems and suggesting concise fixes.
• Third, you will test voice recognition output by speaking prescribed prompts, noting transcription errors, pronunciation gaps, and language-variant issues.
I will supply detailed guidelines, evaluation rubrics, and annotation tools; you simply log in, follow the task queue, and record findings inside the platform. Because experience with search engines, common AI tools, and prior evaluation work is required, you should already be comfortable working with rating interfaces, browser extensions, and ticketing systems such as Jira or Asana.
Deliverables for each weekly sprint
1. Completed rating batches with 95 %+ internal agreement.
2. A short retrospective report summarizing patterns, blockers, and recommendations.
3. Verified voice-test recordings along with transcription error logs.
Work begins with a paid onboarding set to confirm guideline mastery. Steady weekly volume follows once quality is proven.
• First, you will rate the relevance of live web search results against real user queries, flagging mismatches and edge cases.
• Second, you will review AI-generated snippets, answers, and summaries, highlighting factual errors, bias, or tone problems and suggesting concise fixes.
• Third, you will test voice recognition output by speaking prescribed prompts, noting transcription errors, pronunciation gaps, and language-variant issues.
I will supply detailed guidelines, evaluation rubrics, and annotation tools; you simply log in, follow the task queue, and record findings inside the platform. Because experience with search engines, common AI tools, and prior evaluation work is required, you should already be comfortable working with rating interfaces, browser extensions, and ticketing systems such as Jira or Asana.
Deliverables for each weekly sprint
1. Completed rating batches with 95 %+ internal agreement.
2. A short retrospective report summarizing patterns, blockers, and recommendations.
3. Verified voice-test recordings along with transcription error logs.
Work begins with a paid onboarding set to confirm guideline mastery. Steady weekly volume follows once quality is proven.