Live AI Conversation Evaluator: JSON & Portuguese

Lifted (Upwork); United States; Remote; USD 34.000 - 83.000; Teilzeit

Benefits dieser Stelle

Flexible schedule

Remote work

Zusammenfassung

Lifted is seeking experts to engage in live conversations with AI systems, review and label transcripts, and provide written justifications based on rubrics. You will assess language quality, technical accuracy, and JSON formatting while capturing session artifacts and ensuring proper environment setup.

Ideal candidates demonstrate strong Brazilian Portuguese-language fluency and JSON expertise, with flexible remote scheduling for a 2–3 month project that may extend with continued opportunities.

Qualifikationen

• Must pass a basic computational concepts assessment (JSON structures, function, general reasoning, and language proficiency) in about 15 minutes.
• Operator requirements include high-level spoken language skills (C1/C2) in the evaluation language and JSON literacy.
• Audio setup must support clear, noise-free recordings and reliable performance.

Aufgaben

• Participate in live, real-time conversational scenarios with AI systems.
• Review and label conversation transcripts.
• Evaluate AI-generated responses using structured rubrics with written justifications.
• Assess responses for language quality, customer experience, technical accuracy, formatting, and JSON structure.
• Capture session artifacts, including recordings, Session IDs, and tool-call/console logs.
• Prepare environment and verify tools before each session.

Kenntnisse

JSON fluency

Language proficiency

Attention to detail

Jobbeschreibung

Lifted is seeking experts to engage in live conversations with AI systems, review and label transcripts, and provide written justifications based on rubrics. You will assess language quality, technical accuracy, and JSON formatting while capturing session artifacts and ensuring proper environment setup.

Ideal candidates demonstrate strong Brazilian Portuguese-language fluency and JSON expertise, with flexible remote scheduling for a 2–3 month project that may extend with continued opportunities.