Remote AI QA Evaluator: Customer Support & LLM Benchmark

Innodata Inc.; España; Presencial; EUR 21.000 - 34.000; Jornada completa

Descripción de la vacante

Innodata Inc. is seeking detail-oriented evaluators to conduct human quality evaluations for an enterprise AI customer support product on social media.

This remote, freelance role requires strong English, sharp analytical skills, and the ability to follow complex guidelines. You will evaluate AI responses using provided knowledge bases, participate in dual reviews and daily calibration, and maintain productivity targets with an average handling time of about 45 minutes per job.

Formación

• Experience in customer service, call centers, retail, or customer communications.
• Excellent English proficiency with strong grammar and comprehension.
• Strong analytical and attention-to-detail skills.
• Ability to follow complex guidelines consistently.
• Comfortable using web-based evaluation and labeling tools.
• Availability for 6 hours per day.
• Contract: freelance; duration: 2 months (extendable).

Responsabilidades

• Evaluate AI-generated customer interactions against quality rubrics.
• Verify AI responses against FAQs, product catalogs, SOPs, and other approved sources.
• Participate in dual reviews and daily calibration to maintain evaluation consistency.
• Follow detailed rubrics and logic trees while meeting productivity targets.
• Maintain an average handling time of approximately 45 minutes per job.

Conocimientos

Customer service

Attention to detail

Analytical skills

Guideline compliance

Web-based tools

Herramientas

Labeling tools

Descripción del empleo

Innodata Inc. is seeking detail-oriented evaluators to conduct human quality evaluations for an enterprise AI customer support product on social media.

This remote, freelance role requires strong English, sharp analytical skills, and the ability to follow complex guidelines. You will evaluate AI responses using provided knowledge bases, participate in dual reviews and daily calibration, and maintain productivity targets with an average handling time of about 45 minutes per job.