Job description
Tasks include:
Participating in live, real-time conversational scenarios with AI systems
Reviewing and labeling conversation transcripts
Evaluating AI-generated responses using structured rubrics and providing written justifications
Assessing responses for language quality, customer experience, technical accuracy, formatting, and JSON structure
Capturing required session artifacts, including recordings, Session IDs, and tool-call/console logs
Environment preparation and tool verification before each session
Experts should speak naturally and spontaneously, as if responding in real time, without pre-planned, written, or scripted responses - the client is evaluating natural voice cadence.
To be considered for this project, experts are required to pass a basic computational concepts assessment (JSON structures, function, general reasoning, and language proficiency), which takes about 15 minutes to complete, and will not be compensated. If the expert passes the assessment, they will be provided with further onboarding steps.
“Operator” requirements:
Strong spoken-language proficiency: C1/C2-level fluency (CEFR) in the target evaluation language, verified through a speaking/listening assessment, not just a resume claim.
Technical literacy: JSON fluency required - reading, validating, troubleshooting, and identifying structural/formatting issues.
Audio setup: capable of producing clear audio quality with no background noise.
Ideal experts will have strong attention to detail, strong written communication skills, and the ability to evaluate both the customer experience and the technical correctness of a response. A solid understanding of JSON is especially important, including the ability to read, validate, troubleshoot, and identify structural or formatting issues.
An Enterprise client is seeking Welsh-speaking coding experts with JSON expertise to combine language fluency, customer service judgment, and technical evaluation skills to help improve the quality, accuracy, and usability of AI-generated outputs, in an audio format.
Work is expected to begin right away, and continue for 2-3 months, with the potential for additional work beyond this initial project.
Experts may work at any schedule of their choosing, for as long as tasks are available. Experts (Operators) will review a mix of customer service-style responses and technical content, with a strong focus on coding knowledge, JSON structure, formatting accuracy, and overall response quality. Experts should be comfortable identifying whether outputs are clear, helpful, natural, technically accurate, and aligned with project guidelines.
Fluency in Welsh, as spoken in the United Kingdom, is required. Experts with multilingual fluency may be offered work for this project and/or others, in additional languages.
Originally posted on Himalayas
Who can apply
Eligible countries: United States. Accepted UTC offsets: UTC-10, UTC-9, UTC-8, UTC-7, UTC-6, UTC-5, UTC+14. Review the full description for employer-specific work authorization, residency and schedule requirements.