ML Engineer
Neurostevye Tekhnologii LLC
Nov 2025 — PresentBuilt MCP and tool use systems, improved insurance recommendation inference time by 60%, and delivered customer profiling and vector search tools.
• Designed an ecosystem of MCP servers using STDIO and HTTP for LLM agents to interact automatically with knowledge bases, FAQs, and file systems.
• Implemented tool use infrastructure and a full stack React interface for managing collections, prompts, and data synchronization.
• Modernized an insurance product recommendation system and automated inference, reducing weekly output time by 60%.
• Automated the Profiling project to generate customer profiles and calculate business metrics, including renewals and survival rates.
Teacher / Programmer
Phystech College
Nov 2023 — Nov 2025Taught ML and Python, deployed a RAG admissions assistant that automated 70% of answers, and prepared students for competition placements.
• Developed and deployed an LLM based RAG question answering system that automated 70% of admissions office answers.
• Developed an educational data infrastructure based on Delta Lake and a model for summarizing problem domains.
• Taught ML and Python and prepared students who earned more than eight placements in Professional and WorldSkills competitions in ML, Big Data, and robotics using AI.
Data Scientist
Stelspoddon LLC
Jun 2023 — PresentDeveloped computer vision product counting and an NLP email classifier integrated with 1C to improve customer communications.
• Developed an experimental computer vision system for counting products.
• Created an automated email classification system using NLP, integrated with 1C, to improve customer communications efficiency.
Data Scientist
Samolet Group
A major Russian property developer and homebuilder.
Aug 2024 — Nov 2024Prepared computer vision data and trained BERT and RoBERTa classification models for NLP tasks including NER and summarization.
• Performed data labeling with CVAT, automation, preprocessing, and data analysis for computer vision projects.
• Developed and trained multilabel and binary classification models based on transformer architectures such as BERT and RoBERTa for NLP tasks including NER and summarization.
ML Developer
Project for the Government of Moscow Region
Dec 2022 — Mar 2023Led deployment of an OpenCV and YOLOv5 system for visual food quality checks in schools.
• Led development and deployment of a visual food quality inspection system for schools using OpenCV and YOLOv5. The project reached deployment.