AI Quality Analyst (Personalization): German

What you'll do

German is one of the larger markets for this Gemini personalization project, and Turing wants analysts who can judge whether the model's German replies use personal context well. You design short conversations (one to five turns) that depend on your own history, then evaluate the answers for grounding, natural integration and helpfulness.

Expect side-by-side comparisons of two model responses, a written justification tied to specific turns, a check of "Debug Info" to confirm what data the model really used, and a clean-up pass that deletes your evaluation chats.

Who fits

• Strong written and reading German, the focus language
• Ready to use your real, primary Google account with personal data sources enabled
• A good eye for subtle naturalness differences and inferences that are not supported
• Clear, structured rationales
• Degree or equivalent in a relevant field (policy, law, ethics, linguistics, journalism, computer science and similar); annotation or content moderation experience preferred
• Desktop or laptop and reliable internet

What it pays

$15/hour per the ad, for a three-month engagement. The Dutch and Polish versions pay $20/hour but last only one month, so German sits with the longer, lower-rate group.

Worth knowing

Good:

• Three months of potential work, with 4 to 40 hours a week
• Clearly defined task with a stated rubric (grounding, integration, helpfulness)

Less good:

• It uses your actual personal data, which not everyone is comfortable with
• Full-time local availability and 4 hours of PST overlap are expected, which is a late shift for Europe
• $15/hour is modest for a language where analytical skill is required
• Contractor status, no benefits

How hiring works

A Job Interest Form, then a 24-hour assessment, then a pre-onboarding conversation for shortlisted candidates.

About this listing

Turing's own rate, hours and length; the comparison is ours. See Turing.