AI Quality Analyst (Personalization): Vietnamese

What you'll do

Turing is building out coverage of Gemini's personalization feature in Vietnamese. Your job is to act as a real user: write one-to-five-turn prompts that depend on your own life and history, then audit the replies. Did the model assert something about you that its data does not support? Did it blend your details in naturally or announce them awkwardly? Was it helpful?

You also rank pairs of responses side by side, justify each ranking in writing with turn references, check "Debug Info" for the data sources used, and delete the test chats afterwards.

Who fits

• Strong reading and writing in Vietnamese, the project's focus language
• A primary personal Google account you are happy to use with personal data sources enabled
• An eye for tone: Vietnamese pronoun and register choices can make a reply feel unnatural, which is the kind of issue this task asks you to catch (our observation, not the ad's)
• Clear written rationales
• Relevant degree or equivalent experience; annotation or AI evaluation background strongly preferred
• Desktop or laptop with a stable connection

What it pays

$15/hour per the ad, for three months. The Dutch and Polish versions pay more ($20) but last one month.

Worth knowing

Good:

• Three months of work, flexible between 4 and 40 hours a week
• Fewer applicants are likely for Vietnamese than for major European languages (our guess)

Less good:

• Your own emails and activity are used for testing
• Full-time availability in your local time zone and 4 hours of PST overlap are expected, which means a night shift in Vietnam
• Lower of the two pay tiers
• Contractor status

How hiring works

Interest form, then a 24-hour assessment, then pre-onboarding discussion.

About this listing

Turing's text for rate and terms; the extra commentary is ours. See Turing.