At Welodata, I evaluated AI-generated text and code for accuracy, structure, and cultural fit, then rewrote responses to improve language quality and model alignment. I also developed prompts and responses for conversational training datasets and RLHF pipelines.
My projects include a RAG chatbot for domain-specific questions, an OCR pipeline for extracting data from documents, and AI agents for market research and KPI extraction. I also developed an n8n workflow that combines APIs and LLM summarization to generate and distribute newsletters.
