Training Turk

Cantonese Language Evaluator

$30–40/hr · micro1

You evaluate whether AI systems generate accurate, natural, contextually appropriate Cantonese text and speech, identifying linguistic errors and improvement opportunities.

What you would do

  • Assess AI-generated Cantonese for grammatical correctness, natural fluency, and appropriate register for the context
  • Identify linguistic errors: syntax mistakes, vocabulary misuse, tone issues, regional differences, or cultural misunderstandings
  • Annotate language data and assign quality labels supporting AI model development
  • Escalate complex linguistic cases requiring deeper analysis or specialist judgment
  • Document evaluation findings and insights to refine processes

Who they want

  • Fluent mastery of Cantonese with ability to communicate naturally in both formal and casual registers
  • Hands-on background in language evaluation, translation work, linguistic annotation, or language technology initiatives
  • Outstanding communication abilities in spoken and written form with meticulous focus on language detail
  • Experience with content review, transcription work, or accuracy assessment projects
  • Analytical mindset and ability to apply consistent criteria to diverse language samples

Main skills

Cantonese

What the interview asks about

  1. 1.Detecting unnatural but grammatically correct output

    AI can produce syntax that passes grammatical rules but violates native speaker intuition; you identify where output sounds 'wrong' even when you cannot immediately say why.

    For example: “An AI generated a customer service response using formal literary Cantonese in a context where colloquial speech is expected. Describe what's linguistically correct but pragmatically wrong, and how you'd flag this.”

  2. 2.Recognizing register and formality mismatches

    Cantonese varies sharply by social context and speaker relationship; AI often fails to calibrate tone and formality appropriately to audience.

    For example: “A model generated a message to a supervisor using the casual second-person pronoun when formal address is expected. What makes this a linguistic error rather than just cultural misstep?”

  3. 3.Identifying lexical and colloquial confusion

    Homophone and synonym confusion is common in speech-to-text; you spot when word choice creates ambiguity or unintended meaning.

    For example: “An AI transcribed spoken Cantonese where a homophone rendered as the wrong written character, changing 'prosperous business' to an insulting phrase. How would you categorize and document this error?”

  4. 4.Evaluating translation adequacy

    Concepts don't map one-to-one across languages; you judge whether AI translation preserves meaning or collapses nuance in translation from English or other languages.

    For example: “A model translated 'I'm cold' from English into a Cantonese form that means 'I am feeling lonely or isolated,' not 'I feel low temperature.' What's the linguistic distinction you'd explain in feedback?”

A task you may get

Find an example of Cantonese you encounter (speech transcript, written message, translation) that contains a linguistic error you would flag: describe the error, why it's wrong, and what the corrected version should be.

How to prepare

  • Collect 3-4 examples of common Cantonese linguistic errors: homophones misused, formality mismatches, regional usage confusion
  • Prepare examples of Cantonese that is grammatically correct but sounds unnatural, explaining why native speakers would notice
  • Review terminology for linguistic categories (syntax, lexicon, pragmatics, register) you'll use in feedback
  • Recall translation challenges you've encountered between English and Cantonese

The facts

Pay
$30–40/hr
Open to
Bangladesh, Hong Kong, India, Indonesia, Japan, Kazakhstan, Kyrgyzstan, Malaysia, Pakistan, Philippines, Singapore, Sri Lanka, Taiwan, Thailand, Uzbekistan, Vietnam, Austria, Belarus, Belgium, Denmark, France, Germany, Greece, Italy, Netherlands, Portugal, Russia, Spain, Switzerland, United Kingdom, Argentina, Brazil, Chile, Colombia, Mexico, Peru, Algeria, Bahrain, Egypt, Iraq, Jordan, Kuwait, Lebanon, Libya, Morocco, Oman, Palestine, Qatar, Saudi Arabia, Tunisia, United Arab Emirates, United States, Canada, Nigeria, Kenya, South Africa, Ghana, Ethiopia
Field
Language Audio
Role type
Generalist
Posted
9/11/2026
Places left
8

We wrote this page from the public micro1 listing. It may be out of date, so read the full posting before you apply.