Bilingual Cantonese Psychiatrist
$100-200
/hr
Native and fluent Cantonese speakers (from Hong Kong, Guangdong, Macau, and overseas communities) are in high demand for AI training. Cantonese is a complex, tonal language underrepresented in AI training data, commanding premium rates. Tasks include voice recording, translation evaluation, tonal accuracy assessment, accent evaluation, and cultural context checking. Pay: $24-$50/hr depending on task type and specialization. Fully remote and flexible scheduling.
Part of: Chinese
41 Jobs
Page 1 of 2
$30-40

Trending jobs + new guides + real pay data. Every Tuesday.
$8-70
/hr
$40-60
/hr
$25-35
$10-35
/hr
$15-25
/hr
$15-25
$40-60
/hr
$8-70
/hr
$8-70
/hr
$8-70
/hr
$6-70
/hr
$8-70
/hr
$8-70
/hr
One browser alert a day when new roles land, Cantonese first. No email needed.
Cantonese evaluation tasks address a language that AI models handle poorly relative to Mandarin, making expert review especially valuable. You'll evaluate AI Cantonese for correct use of Cantonese-specific vocabulary (distinct from Mandarin), accurate tonal representation (Cantonese has 6-9 tones depending on analysis, compared to Mandarin's 4), and authentic Cantonese sentence structure, Cantonese grammar differs from Mandarin in aspect markers, sentence-final particles, and topic-comment constructions. Written Cantonese tasks assess whether AI correctly uses colloquial written Cantonese characters. Audio tasks involve recording in authentic Cantonese (Hong Kong, Guangzhou, or overseas variety as specified). Native speaker or near-native fluency is required. Cantonese expertise is rare in AI training datasets and commands premium compensation.
Less. Mandarin roles outnumber Cantonese ones on every platform here, and Cantonese listings lean toward voice, transcription and tonal-accuracy work rather than text evaluation. Fewer native speakers apply, so move quickly on a listing that names Cantonese.
Voice recording, translation evaluation, transcription, tonal accuracy assessment, accent/dialect evaluation, and cultural context checking. Voice roles generally offer higher rates than text-based tasks.
For written tasks, usually yes: Cantonese projects most often use Traditional characters as written in Hong Kong. Voice and transcription roles care about spoken fluency and tone accuracy first.
This page is showing 41 listings, sourced from 9 platforms: Mercor (15), Meridial (7), Micro1 (4), rws (4) and Turing (4). The most recent one was picked up today. The count moves as platforms post and close roles, so treat it as a snapshot of today rather than a fixed figure.
Most sit between $24 and $50 an hour, with a median of $45, measured across the 33 listings on this page that quote an hourly rate. The highest rate currently posted here is $200 an hour. These are the figures the platforms published, not our estimates, and the rate you are offered depends on how you do in that platform's assessment.
Both appear here. 1 listing pays per task or per project rather than by the hour, and the rest quote an hourly rate. Per-task work suits people who are fast and accurate, since your effective hourly rate depends on how long each unit takes you.
Mercor (15), Meridial (7), Micro1 (4), rws (4) and Turing (4). No single one dominates this category, so it is worth holding accounts on more than one. Most contracts allow you to work through several platforms at the same time.
The roles here fall across 10 areas, led by Languages (22), Data Science (4), Software Engineering (3) and QA Testing (2). The requirements named most often are Cantonese, AI Training, Style Guides, English and bilingual communication. Each listing spells out its own task type, so read the individual page before applying.
Browse all job categories and find your next AI training opportunity.
View All Categories