I am an AI / LLM Engineer and Data Scientist with 3+ years of industry experience building and deploying production-grade machine learning systems, primarily in NLP and LLM-based applications.
I work at the intersection of classical ML, modern LLM pipelines, and applied MLOps, focusing on solving real business problems with measurable impact. I have hands-on experience delivering end-to-end ML solutions — from problem formulation and experimentation to online inference in production.
LLM & NLP
- Prompt engineering and LLM-based system design
- Retrieval-Augmented Generation (RAG) pipelines
- Embedding models and semantic search
- Text classification and Named Entity Recognition (NER)
- Practical experience with Russian-language NLP and multilingual models
Models & Frameworks
- BERT, RoBERTa
- LLaMA, Mistral, Qwen, DeepSeek
- PyTorch, Hugging Face, scikit-learn
Production & MLOps
- Online inference and API-based ML services
- FastAPI-based model serving
- Docker, CI/CD pipelines
- Experiment tracking with MLflow
CheckDocAI AI-powered Telegram bot for automated document quality control, successfully deployed in commercial use.
- Led a small cross-functional team (data scientists + backend engineer)
- Designed and implemented the ML pipeline (CV + NLP components)
- Achieved a measurable business impact: ~40 human-hours saved per month
- Tech stack: PyTorch, YOLO, ONNX, Albumentations, CVAT, aiogram
- Contributor to PySAD — an open-source library for anomaly detection in streaming data
- Author of technical articles on anomaly detection methods (HBOS, ECOD, Isolation Forest)
- Speaker at Positive Hack Days 2025: “Anomaly Detection with Python”
- Specialist degree in Thermal Engineering, Moscow Aviation Institute
- Specialist degree in Linguistics and Translation, Moscow Aviation Institute
- Solid foundation in mathematics, statistics, and algorithms