Speech AI Evaluation Specialist

About the role

This role places you at the intersection of artificial intelligence and linguistic evaluation, where your judgment directly influences the quality of AI-generated Vietnamese content. You will own the end-to-end evaluation cycle for speech interactions, determining whether outputs meet standards for relevance, accuracy, clarity, and overall quality. The position requires a high degree of autonomy, as you will interpret project guidelines independently and apply them consistently across diverse conversations. You are responsible for providing objective ratings and detailed feedback that accurately reflect both technical performance and user experience. This contract represents a direct opportunity to shape the development and safety of contemporary AI models through meticulous human oversight. By adhering to established quality and productivity expectations, you ensure the reliability of the evaluation data. Your work will contribute to building more robust and user-friendly AI systems that serve Vietnamese-speaking communities. This engagement is designed for individuals who view evaluation not as a mechanical task, but as a critical analytical function.

Key facts

What you'll do

Conduct short voice conversations with AI models using a designated web-based platform.

Follow specific scenarios and prompts provided by the project framework to guide each interaction.

Evaluate AI-generated responses against a multi-dimensional set of criteria including relevance, accuracy, clarity, and overall quality.

Assign objective ratings and construct written feedback that documents the reasoning behind each assessment.

Monitor your own performance metrics to ensure adherence to established quality and productivity expectations.

Apply project guidelines with consistency to maintain data integrity across a high volume of conversational samples.

Classify interactions based on predefined quality dimensions to support downstream analysis.

Identify and document instances where AI responses fail to meet clarity, accuracy, or relevance standards.

Maintain a detailed understanding of evaluation criteria to reduce ambiguity in scoring decisions.

Support the iterative improvement of the AI model by providing reliable human feedback loops.

Adapt to evolving project requirements by quickly learning updated evaluation protocols and instruction sets.

Ensure that all interactions remain confidential and are handled in strict compliance with data security policies.

Requirements

You must be a native-level speaker of Vietnamese to accurately assess linguistic nuances and cultural appropriateness.

You must possess fluent or advanced proficiency in English, demonstrated at levels B2 through C2 on the Common European Framework of Reference for Languages.

You must be a resident of Malaysia, as this contract is specifically targeted at nationals or long-term residents of the country.

You must be able to work remotely from home, requiring a stable internet connection and a suitable environment for focused evaluation.

You must be available to commit a minimum of 10 hours per week to the project on a part-time basis.

You must be flexible regarding work hours, accommodating a schedule that allows you to work whenever you are able within the platform's operational window.

You must be able to start immediately upon acceptance of the contract terms.

You must treat the role with the seriousness of a professional engagement, meeting deadlines and quality standards consistently.

Nice to have

Prior experience with machine learning tasks, such as data collection, preprocessing, or evaluation, is valued.

Experience in data annotation and labeling, particularly for speech or language datasets, is considered an asset.

Understanding of data quality assurance principles will enhance your ability to identify inconsistencies or errors in AI output.

Practical notes

This is a freelance position with a flexible schedule, requiring 10 or more hours of work per week.

The role is fully remote and work-from-home, eliminating the need for physical office attendance.

Compensation is offered at a rate of 18 USD per hour, with timely payments arranged by the platform.

The engagement is ideal for students, part-time workers, stay-at-home parents, and gig workers seeking supplemental income.

There are no specific travel requirements or visa sponsorship considerations for this role.

The start date is immediate, though the duration is to be confirmed based on project needs and mutual agreement.

#LI-PR15