Work type
Model evaluation
Evaluation sets that test what a model knows about the region: its languages, laws, places and everyday life.
Get a pilotTypical tasks
- Eval prompts written with reference answers
- Fact checks that need local knowledge
- Regression runs on each new model version
Who does it
Tier 2 evaluators; domain experts for specialist sets.
Managed rates
- Kazakh, Uzbek, Kyrgyz, Tajik, Mongolian native QA and alignment $24–30 /h
- Serbian, Bosnian, Macedonian, Bulgarian, Albanian native evaluator $18–22 /h
- Domain expert: engineering, medicine, law, finance $40–55 /h
Languages
Pilot
Start with a paid two-week pilot
10 vetted workers in your tool, ready within 10 business days (5 for Russian). If a managed team’s weekly gold-set agreement falls below 90%, the hours below it aren’t billed.