AI Trainer & Data Contributor
As an AI Trainer & Data Contributor at Toloka, I performed search relevance evaluation and fine-tuned large language models through data labeling. My work included conducting side-by-side (SbS) comparisons, contributing to reinforcement learning from human feedback (RLHF), and ensuring quality via regular audits. I classified and analyzed text data to support improvements in search algorithms and conversational AI models. • Deep-dive search query analysis for search engine ranking improvements • Side-by-side LLM response comparisons for RLHF • Consistency checks and "Golden Set" audit success • Linguistic text analysis for sentiment, tone, and grammatical accuracy