Lightly

-

Posted on: 14 September, 2025

Working Student, AI Training / Model Evaluation

Evaluate and refine AI models and their outputs to improve reliability and performance of AI systems.
Prompt-Response Evaluation: assess if AI responses accurately address prompts, follow instructions, and meet quality standards.
Language & Context Check: judge AI understanding of English, including grammar, tone, and nuance handling.
Domain Relevance: verify responses fit within the expected domain (technical, general knowledge, conversational).
Instruction Following: identify deviations from prompts or additions of irrelevant information.
Optional visual checks: when tasks involve visual outputs, check image accuracy and caption clarity depending on project needs.

Tags:

ai

ml

Share the job: