Summary
Artem Zabolotnyi is a Senior ML Researcher with 9 years of experience combining deep learning research and full-stack engineering, currently driving transformer-based risk modeling at SberBank. He has led development and deployment of large-scale transformer models (10–30+ billion parameters) across clusters of hundreds of GPUs, producing improvements in credit scoring that translated to tens of billions of rubles in financial impact. His background spans CV, NLP, and audio, plus practical expertise in PyTorch, DeepSpeed, MLflow, Hadoop/Spark and production ML pipelines, enabling seamless research-to-production workflows. Earlier full-stack roles gave him hands-on skills in microservices, CI/CD and frontend engineering, so he comfortably bridges model research and service engineering. As a PhD researcher and lecturer at Skoltech he explores uncertainty estimation and efficient transformer designs, and he enjoys mentoring peers and distilling scientific papers into practical architectures. An unassuming detail: he pioneered using large GPU clusters at SberBank and won the 2023 R&D Breakthrough prize for the X-Former project.
9 years of coding experience
3 years of employment as a software developer
Bachelor's degree, Cyber Security, Bachelor's degree, Cyber Security at Санкт-Петербургский Государственный Университет Аэрокосмического Приборостроения
Doctor of Philosophy - PhD, Artificial Intelligence, Doctor of Philosophy - PhD, Artificial Intelligence at Skolkovo Institute of Science and Technology