Education & Career
My professional journey
Work Experience
Full-Stack Developer
Freelance
Commercial projects end to end — from the brief to a running production with monitoring. Flagship: TicketHub, an open-source PHP 8.4 ticketing system — a ~45-endpoint REST API, an OWASP-aligned security layer, and Kanban N+1 cut from 201 queries to 2 per page.
Web Developer / System Administrator
Russian State Children's Library (RGDB)
I maintain and extend the library's PHP web services and rolled out CI/CD on GitHub Actions — deploys went from 2 hours to 15 minutes. I own the 99.9% SLA for the server infrastructure (Prometheus + Grafana) and lead a team of two system administrators.
Junior System Administrator
Russian State Children's Library (RGDB)
L1-L2 support for 100+ workstations (Windows, Linux, macOS). Average ticket close time — 2 hours, 95% user satisfaction. Built a knowledge base of 50+ articles that cut senior escalations by 30%. First Python/Bash scripts here — saved the team 2-3 hours daily. Fixed ~30 bugs in legacy PHP web services and sped up the main site by 20%.
Education
Moscow Witte University
Bachelor's Degree, Applied Informatics — AI & Data Analysis
Studied alongside work. Focus on machine learning, data engineering, and applied AI. Coursework: predictive maintenance for a metro-train air compressor on the MetroPT-3 dataset (ROC-AUC 0.993). The research on ensemble methods for cardiovascular disease diagnosis was published in a Russian HAC-listed journal (MOIT, category K2, 2026).
Projects during studies
- →Heart Disease ML BenchmarkMulticenter ML benchmark on 1,904 patients from 6 heart-disease datasets — CatBoost wins at 0.948 ROC-AUC
- →MetroPT Predictive MaintenanceEarly detection of air leaks in a metro-train compressor 30 minutes before failure — 1.5M sensor readings, 16 models compared, tuned GradientBoosting at 0.993 ROC-AUC
- →GenreNeuroNeural classifier for Russian children's books — 89.16% accuracy across 6 genres from text descriptions
- →Children Book DatasetPython scraper that builds a labeled dataset of Russian children's books — made to feed the GenreNeuro classifier
MGOK (Moscow State College)
Associate's Degree (SPO), Network & System Administration
Where it all started. Learned networking, Windows Server, Linux, and wrote my first Bash scripts. Got hooked on automation — if something took more than 10 minutes by hand, I'd script it.
Certificates
Systems Analysis
Karpov Courses · Jan 2026
Publications
Ensemble Machine Learning Methods for Predictive Diagnostics of Cardiovascular Diseases: Comparative Analysis on a Multi-Center Dataset
Modeling, Optimization and Information Technology (MOIT), 2026, Vol. 14, No. 6(57) · Russian HAC list, category K2
Eight machine-learning algorithms compared on a combined multi-center sample from six databases (n = 1,904). CatBoost reached ROC-AUC 0.948 [0.922–0.966] and F1 0.884 with the best calibration; an ablation study showed that seven features retain 97.5% of full-model performance. Pairwise comparison used DeLong and McNemar tests with Bonferroni correction. Co-authors: D.I. Veselov, N.A. Andriyanov.
DOI: 10.26102/2310-6018/2026.57.6.017