In this role, you will perform A/B preference testing of machine-generated translations from English into Indic languages, helping to evaluate and improve the quality of AI-generated content.
Each sample is reviewed by three independent validators working in a blind process.
Responsibilities
As a Translation Validator, you will independently review an English source document alongside two machine-generated translations (Variant A and Variant B), select the preferred translation, and provide a written justification for your choice.
Review English source documents alongside two machine-generated Punjabi translations.
Evaluate both variants based on accuracy, fluency, and overall translation quality.
Select the preferred translation and provide a clear written justification for your assessment.
Complete assigned samples independently and within established timelines.
Adhere strictly to project and client guidelines.
Required skills
Native-level fluency in Punjabi (written and spoken).
Strong written communication skills in the target language.
Proficiency in English.
Ability to work independently with high attention to detail and consistency.
Preferred skills
Background in linguistics, translation, or language quality assessment preferred.
Familiarity with Romanized script conventions for the relevant language is a plus (for transliteration tasks).
Thriveth makes AI data-training work easier to find, understand, and navigate. We replace uncertainty with clear opportunities, realistic expectations, and insights from real application journeys.
Hiring companyWelo Data
Turing