Freelance AI Training and LLM Evaluation (Stellar AI, Toloka AI), 2025 - Current
TextFreelance AI training and LLM evaluation work on the Stellar AI and Toloka AI platforms. I evaluated and ranked AI-generated responses using structured rubrics across factual accuracy, relevance, reasoning quality, completeness, clarity, and instruction adherence. I performed comparative assessment by selecting the stronger output using evidence-based judgment and guideline compliance. • Evaluated and ranked responses for quality, accuracy, and relevance using criteria and changing task guidelines • Conducted fact-checking and online research to identify unsupported claims, factual inaccuracies, and logical weaknesses • Completed Greek-language evaluations requiring native-level understanding of grammar, meaning, tone, fluency, and cultural context • Covered business/management, finance/economics, logic, research, and general STEM subject areas, maintaining consistency across repeated tasks