Project BVSR: European Spanish (es-ES) ←→ Basque (eu-ES)
Budget: $10 USD
ABOUT THE PROJECT
The Vega project is a bilingual AI data initiative producing high-quality code-switching conversation scripts — dialogues that naturally alternate between two languages, simulating real-world call centre interactions between agents and customers.
As a Script Reviewer, you will conduct secondary manual QA on scripts of 150–200 lines each, cross-checking machine-applied corrections and resolving any remaining objective errors before final delivery.
WHAT YOU'LL BE DOING
-Reviewing bilingual code-switching call centre scripts for accuracy and consistency
-Verifying that machine-applied (Global Check) corrections have been properly implemented
-Identifying and fixing objective errors: grammar, spelling, morphology, and word formation
-Correcting formatting issues including spacing, punctuation, and XML/language tag errors
-Resolving factual inaccuracies such as invalid postcodes or unnatural email/IBAN formats
-Removing cross-language repetition and correcting non-standard hybrid word forms
-Maintaining conversation coherence and flow after any modifications
-Each script typically takes 15–25 minutes to review
WHAT WE"RE LOOKING FOR
Native or near-native fluency in English and at least one of the listed target languages
Strong attention to detail and comfort working with structured, tagged text formats
Familiarity with XML-style language tags (e.g. [EN]…[/EN], [ES]…[/ES])
Ability to distinguish objective errors from subjective style preferences — fix the former only
Reliable internet connection and ability to work independently via our client's online platform
Experience in linguistics, translation, or language QA is a plus but not required
IMPORTANT GUIDELINES
Fix only clear, objective, verifiable errors. Do not rewrite for naturalness or style — preserve the original script content and structure at all times. If an error is ambiguous or subjective, leave it untouched. AI tools such as ChatGPT or Gemini may be used for reference only; all final decisions must be your own.
The Vega project is a bilingual AI data initiative producing high-quality code-switching conversation scripts — dialogues that naturally alternate between two languages, simulating real-world call centre interactions between agents and customers.
As a Script Reviewer, you will conduct secondary manual QA on scripts of 150–200 lines each, cross-checking machine-applied corrections and resolving any remaining objective errors before final delivery.
WHAT YOU'LL BE DOING
-Reviewing bilingual code-switching call centre scripts for accuracy and consistency
-Verifying that machine-applied (Global Check) corrections have been properly implemented
-Identifying and fixing objective errors: grammar, spelling, morphology, and word formation
-Correcting formatting issues including spacing, punctuation, and XML/language tag errors
-Resolving factual inaccuracies such as invalid postcodes or unnatural email/IBAN formats
-Removing cross-language repetition and correcting non-standard hybrid word forms
-Maintaining conversation coherence and flow after any modifications
-Each script typically takes 15–25 minutes to review
WHAT WE"RE LOOKING FOR
Native or near-native fluency in English and at least one of the listed target languages
Strong attention to detail and comfort working with structured, tagged text formats
Familiarity with XML-style language tags (e.g. [EN]…[/EN], [ES]…[/ES])
Ability to distinguish objective errors from subjective style preferences — fix the former only
Reliable internet connection and ability to work independently via our client's online platform
Experience in linguistics, translation, or language QA is a plus but not required
IMPORTANT GUIDELINES
Fix only clear, objective, verifiable errors. Do not rewrite for naturalness or style — preserve the original script content and structure at all times. If an error is ambiguous or subjective, leave it untouched. AI tools such as ChatGPT or Gemini may be used for reference only; all final decisions must be your own.