Личные данные скрыты

Volodymyr Shabat
Remote, Ukraine | [открыть контакты](см. выше в блоке «контактная информация»)
LinkedIn | GitHub

Professional Summary
Infrastructure Engineer transitioning to DevOps, currently managing production monitoring and networking for a regional
ISP (50+ nodes). Hands-on daily with Zabbix/Grafana, Linux servers, Cisco/MikroTik networking, and Docker
containerization. Self-host AI inference infrastructure on Proxmox. Bachelor’s thesis: multi-agent NL2SQL system;
published BIRD-UKR, the first Ukrainian-language Text-to-SQL benchmark. Completing EPAM DevOps training
program.

Technical Skills
Monitoring & Observability: Zabbix, Grafana, Prometheus, SNMP, Cacti Weathermap, alerting & runbooks
Systems & Virtualization: Linux (Ubuntu/CentOS/Debian), Windows Server (AD, DNS, DHCP), VMware ESXi,
Proxmox
Containers & CI/CD: Docker, GitLab CI/CD, Nginx, n8n workflow automation
Networking: Cisco IOS, MikroTik, TCP/IP, VPN, VLANs, firewalls, DNS, Wireshark
Databases: PostgreSQL, MS SQL Server, SQL optimization
Scripting & Automation: Bash, Python, PowerShell, SQL

Professional Experience
Infrastructure & DevOps Engineer Kyiv Region, Ukraine
Baryshivka.NET ISP Aug 2024 – Present
– Deployed centralized monitoring stack (Zabbix + Grafana) for 50+ network nodes with custom SNMP
templates, dashboards, and alerting — reduced incident response time by 40%.
– Administer 50+ Cisco/MikroTik devices: routing, VLANs, firewall rules, VPN tunnels.
– Manage production Linux servers (Ubuntu/CentOS): Nginx reverse proxies, PostgreSQL, SSH hardening.
– Maintain virtualized infrastructure on VMware ESXi and Proxmox, provisioning VMs for production and dev
environments.
– Built GitLab CI/CD pipelines and migrated internal services to containerized Docker architecture.
– Deployed n8n workflow automation to orchestrate billing/CRM data synchronization.
– Maintained Cacti Weathermap and LibreNMS for real-time SNMP-driven network topology visualization.

Education & Training
Taras Shevchenko National University of Kyiv Kyiv, Ukraine
Bachelor of Science in Computer Engineering Sep 2021 – Jun 2025
– Thesis: “Multi-Agent Systems for Database Interaction” — architected a multi-agent Text-to-SQL pipeline
replacing GPT-4 with Llama 3.1 70B, achieving 90% cost reduction.
– Created and released BIRD-UKR, the first Ukrainian-language NL2SQL benchmark corpus (8 databases, 3
difficulty tiers).
EPAM DevOps Training Program Remote
Cloud Infrastructure & CI/CD 2025 – Present
Cisco CCNA 3: Enterprise Networking Certification
Routing, switching, network security Sep – Nov 2024

Key Projects
BIRD-UKR — Ukrainian NL2SQL Benchmark | PostgreSQL, Python, LLM-assisted generation
– Designed and built the first comprehensive Ukrainian-language Text-to-SQL dataset: 8 domains, ˜80 tables, 75+
queries across 3 difficulty tiers.
– Implemented reproducible evaluation pipeline with Docker and PostgreSQL for benchmarking NL2SQL models.
– Open-sourced on GitHub: github.com/Leev1tan/BIRD-UKR
Edge AI Inference on Commodity Hardware | Linux, Proxmox, MESA/Vulkan, llama.cpp
– Engineered stable Llama 3 (8B) inference at 32 tok/s on repurposed AMD BC-250 mining GPUs by compiling
custom MESA/RADV drivers and tuning kernel parameters.
 – Self-host AI agent infrastructure on Proxmox with NVIDIA NIM, demonstrating systems-level problem solving.
Enterprise Infrastructure Lab | Windows Server, C#/.NET, Active Directory
– Deployed Windows Server environment: AD DS, DNS, DHCP, IIS, Group Policy in virtualized infrastructure.
– Built C#/.NET automation tools for user provisioning via LDAP; ASP.NET MVC with MS SQL integration.
– Configured routing, firewalls, VPNs on Cisco/Huawei; packet analysis with Wireshark.