In this role, you will review the work submitted by Trainers — pre-seeded questions paired with images and their corresponding "golden" answers — to confirm accuracy, consistency, and adherence to project guidelines.
This is a task-based, remote-friendly opportunity ideal for individuals who are highly meticulous, observant, and skilled at spotting errors or inconsistencies that others may miss.
Responsibilities
Review completed tasks submitted by Trainers, including the original question, the image, and the golden answer provided.
Independently verify whether the golden answer is correct based on the image.
Identify and flag incorrect, incomplete, ambiguous, or inconsistent answers.
Provide clear, constructive feedback or corrections to help maintain overall data quality.
Track recurring error patterns and report them to the project team.
Ensure Trainer output meets the project's quality bar before final sign-off.
Escalate unclear cases (e.g., illegible images, ambiguous questions) to the project team for resolution.
Experience and education
Bachelor's degree required.
Excellent attention to detail and a strong quality mindset.
Good visual observation and reading comprehension skills.
Ability to apply guidelines/rubrics consistently and objectively.
Prior experience in quality assurance, quality control, editing, proofreading, data review, data annotation or data labeling is a plus.
Language requirements
Native/Strong proficiency at Bengali.
Strong written English communication skills, especially for giving feedback.
Equipment requirements
Reliable access to a computer and internet connection.
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