Photo of Nikolai Mozgovoi

Nikolai Mozgovoi

DevOps/DataOps & Analytics Engineer · ML Systems · Ph.D. student

• Rostov-on-Don, Russia

About

AI Researcher and ML Engineer focused on building reliable ML systems and shipping end-to-end data & inference pipelines.I build and support geo-distributed Kubernetes services, data pipelines (Airflow), and analytics/reporting (Superset). Interests: NLP, generative multimodal models, and computer vision.

NLP Computer Vision Kubernetes Airflow PostgreSQL / MSSQL / MySQL Python

Experience

Junior DevOps/DataOps & Analytics Engineer — LLC Commercial Bank RostFinance

Nov 2023 – Dec 2025 · Rostov-on-Don
  • Deployed, administered, and supported services in a geo-distributed Kubernetes cluster.
  • Monitored cluster health in Grafana and performed routine cluster maintenance.
  • Administered PostgreSQL, MSSQL, and MySQL at scale (user management, query optimization, advanced window functions).
  • Designed, implemented, and maintained ETL pipelines in Apache Airflow (full & incremental loads into an enterprise DWH).
  • Implemented data validation and reconciliation with automated anomaly detection, reducing data discrepancies by ~40%.
  • Built dashboards in Apache Superset for reporting and executive decision-making, cutting manual effort by ~75%.
  • Developed proprietary inference optimization reducing latency by 73% vs baseline.
  • Maintained legacy Java microservices (support, bug fixes, stability improvements).
  • Developed internal banking analytics microservices in Python.
  • Built a computer vision system to detect customer visits in bank operation halls.
  • Fine-tuned a wav2vec model for internal banking tasks.

Teaching Assistant — Don State Technical University (DSTU)

Oct 2025 – present · Rostov-on-Don

TA at the Department of Cybersecurity of Information Systems.

Education

Ph.D., Computer Science & Information Processes — DSTU

Sept 2025 – present

Department of Cybersecurity of Information Systems.

M.Sc., Intelligent Media Technologies — DSTU

Sept 2023 – Jul 2025 · GPA 5/5 (Hons)

Thesis: Neural networks for generating 3D animation of humanoid characters.
Advisor: Assoc. Prof. Evgeniy Ladosha.

B.S., Information Systems & Technology — DSTU

Sept 2019 – Jul 2023 · GPA 5/5 (Hons)

Thesis: Creation and Animation of a 3D Character Using Visual Programming (Nodes).
Advisor: Assoc. Prof. Evgeniy Ladosha.

Skills

Languages

Python · SQL · Julia

Infrastructure

Kubernetes · Docker · Git · Ansible · Helm · GitLab CI/CD · ArgoCD · Grafana · RabbitMQ · Redis

Data Tools

Airflow · Superset · dbt

ML

PyTorch · JAX · Triton · ONNX · Transformers · OpenCV · XGBoost · LightGBM · CatBoost

Selected projects

SCADE-Net (Jan 2026)
Single-Center Attention & Defocus-Enhanced Network for generalized deepfake detection (~28k params) improving FaceForensics++ DFD (C23) AUC from 0.945 to 0.977.
DeepfakeComputer VisionECA
SMoG — Semantic Motion Generation (Jan 2025)
Text-to-motion synthesis using CLIP semantics and KAN-enhanced transformers for realistic 3D human motions. Higher R@5, lower MMD and lower FCD vs baseline MotionCLIP.
GenerativeTransformers3D Motion

Other

International experience

Elementary Chinese study (Summer Semester 2024), Harbin Institute of Technology, Harbin, China (July 2024).

Extracurricular

  • AI Hackathon (HSE & Promsvyazbank).
  • Led development of a 3D relief map of Russia (3D model from 2D image, texturing, materials).
  • Lectures on additive technologies and 3D scanning/printing (Tochka Kipeniya, UBiTech; RRAPP initiative).

Last updated: Jan 2026.