Summary
Daniil Vodolazsky is a machine-learning leader with eight years of experience building document understanding and information-extraction systems at Sberbank, now leading a team focused on advanced extraction with LLMs and AI agents. He conceived and delivered GigaQuery, a production framework that extracts structured data from documents without relying on RAG, and helped ship SberJazz videoconference summarization into production. Daniil has driven research-to-product work across AutoNER, multimodal pretraining (IDPFormer on 11M pages), RusDocVQA, and SFT data for GigaChat, combining novel datasets and architectures that outperformed local state of the art. Comfortable mentoring and coordinating cross-team efforts, he pairs hands-on model and pipeline development with business impact—multiple client projects using his tools generated significant revenue. He holds strong academic foundations from NSU, HSE and Stuttgart and contributes to data science education as a TA at Yandex School of Data Analysis.
8 years of coding experience
5 years of employment as a software developer
Data Science, Data Science at Computer Science Center
Master's student Data Science, Master's student Data Science at Higher School of Economics
Computer Science, Computer Science at Yandex School of Data Analysis
Computational Linguistics, Computational Linguistics at University of Stuttgart
Bachelor of Science - BS Applied Mathematics and Informatics, Bachelor of Science - BS Applied Mathematics and Informatics at Novosibirsk State University (NSU)
Русский, English, Украинский