Help shape the future of AI by evaluating the technical accuracy, language quality, and usefulness of AI-generated responses.

Location: Remote
Type: Independent contractor / Part-time
Project duration: up to 3 months

About the role

At Anyone AI, we create high-quality training and evaluation data for frontier AI models.

We’re looking for multilingual experts with coding knowledge and a solid understanding of JSON to evaluate AI-generated technical content and customer service responses.

In this role, you’ll assess whether outputs are clear, helpful, natural, technically accurate, and aligned with project guidelines. Your evaluations will help improve both the customer experience and the technical reliability of AI systems.

What you’ll do

Review AI-generated customer service responses and technical content.
Evaluate responses for clarity, helpfulness, natural language, and overall quality.
Assess technical correctness using your coding knowledge.
Read, validate, and troubleshoot JSON, identifying structural and formatting issues.
Check outputs for formatting accuracy and adherence to project guidelines.
Provide clear, well-reasoned feedback explaining your evaluations.

What we’re looking for

Fluency in one or more of the supported language varieties listed below.
A solid understanding of JSON, including syntax, structure, validation, and troubleshooting.
Coding knowledge sufficient to assess technical content and identify errors.
Strong judgment when evaluating customer service responses.
Excellent attention to detail and strong written communication skills.
Ability to follow detailed evaluation guidelines consistently.

Supported languages/locales

Dutch — Netherlands
Welsh — United Kingdom
French — Canada
German — Germany
Italian — Italy
Portuguese — Brazil
Greek — Greece

Nice to have

Previous experience in AI evaluation, quality assurance, or technical support.
Experience working with APIs and structured data.
A background in customer service, software development, or technical writing.

Voice response instructions

When completing the voice response, speak naturally and spontaneously, as if responding in real time. Do not pre-plan, write, or script your answer—we are evaluating natural voice cadence.

Responses may be up to one minute long. A few sentences are sufficient if that feels natural.

Open to applicants in

Argentina, Uruguay, Chile, Poland, Italy, Czech Republic, Estonia, Ecuador, Portugal, Mexico, Paraguay, Colombia, France, Spain, Peru, Brazil, Germany

AI Coding Evaluator

Začněte e-mailem