Data Scientist & ML Infrastructure Engineer
Building LLM inference infrastructure — from serving optimization on 16×H100 + 8×A100 clusters to platforms for 50–60 researchers — and applied ML: multivariate time series, computer vision, RAG.
I'm a lead data-science expert at Sberbank, where I run inference infrastructure for large language models — DeepSeek, GLM, Qwen, 4B to 400B parameters — on SLURM clusters of 16×H100 + 8×A100, and build tooling used by 50–60 researchers: GitLab, ClearML, LLM gateways, reverse proxies.
Parallel to that: data science at Orb Intelligence (ELT pipelines, labeling ops, vector search at recall@1 = 0.997) and research engineering at the Institute of Precision Mechanics and Control, RAS, where work on satellite telemetry grew into first-year PhD research on multivariate time-series analysis — with a published computer-vision paper on the way.
Outside work: photography and video editing, blogging, powerlifting, playing musical instruments.
Native
B2
Working proficiency — papers, documentation