Timur Khasanov

Built a multilingual conversational AI agent for online market research handling thousands of concurrent sessions, with context management, quality evaluation, and dialogue moderation — doubling insights vs. traditional surveys. Automated extraction and quantification of insights from hundreds of dialogues. Built Talk to Data LLM agent with RAG, tool calling, and sandboxed execution – cutting insight time from hours to minutes. Migrated to unified .NET architecture and built React-based testing tools.

Designed and shipped a suite of products automating data extraction from large sets of open ended answers in online market surveys. Used a variety of NLP algorithms and tools like keyword extraction, embeddings-based search, text classification, sentiment analysis, named entity recognition, and topic modeling — largely as a solo developer

Market research data analysis: Survey data processing, programming, and analysis in SPSS and Python

Master of Science, Faculty of Physics, Molecular Physics department