Skip to content
View mykhailo-kotsar's full-sized avatar

Block or report mykhailo-kotsar

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
mykhailo-kotsar/README.md

Mykhailo Kotsar — DevOps & Infrastructure Engineer

Self-taught infrastructure engineer with 2+ years of hands-on experience in blockchain node operations, cloud infrastructure, Linux administration, and automation. Currently working as a DevOps engineer at a Ukrainian legal-services company, managing its production infrastructure end-to-end, while transitioning toward LLMOps / AI infrastructure.


Stack

OS & Core · Linux (Ubuntu, Debian) · Bash · systemd · Nginx · Docker · docker-compose

Virtualization · Proxmox VE · KVM/QEMU · VM lifecycle management

CI/CD & VCS · GitLab CI/CD · GitHub Actions · Git · Shell scripting

Monitoring · Prometheus · Grafana · Node Exporter · Graylog · Telegram Alerting

Security · SSH hardening · Firewall management (pve-firewall, iptables) · VPN (WireGuard, IKEv2/StrongSwan)

Identity & Access · FreeIPA · Keycloak · WireGuard UI · MDM (Mosyle Business)

IaC & Cloud · Terraform · AWS EC2 · DigitalOcean · Hetzner · Oracle Cloud · Azure

AI & LLMOps · Claude Code (multi-agent architectures, subagents, hooks/guardrails) · Claude & OpenAI APIs · AWS Bedrock · prompt engineering · AI-assisted infrastructure workflows

Web3 · Node Operations (30+ protocols) · Validator setup · RPC infrastructure · DePIN · EVM


Experience

DevOps / Infrastructure Engineer — Agrumiuris (May 2026 – Present) Full infrastructure recovery and administration after previous admin departure.

  • Proxmox VE cluster: 18 VMs, 125GB RAM, 4× NVMe, Hetzner dedicated
  • Security audit: SSH hardening, firewall rules, VPN access control, key rotation
  • GitLab CI/CD: restored offline runner, fixed pipeline configs, set up auto-deploy
  • Backup system: Proxmox vzdump, ZSTD compression, automated rotation (2×/day)
  • MDM: Mosyle Business — 12 corporate MacBooks, DEP enrollment, remote management
  • Monitoring: Grafana + Graylog stack across all VMs
  • Infrastructure: Nginx Proxy Manager (lb-01), PowerDNS (dns-01), WireGuard VPN, FreeIPA, Keycloak, Bitrix24 CRM, AI Agent app (Node.js + OpenAI)

Web3 Infrastructure Engineer — NodeRunner (Mar 2025 – Jan 2026)

  • Operated 30+ blockchain nodes across AWS, Azure, DigitalOcean, Hetzner
  • Built monitoring stack: Prometheus + Grafana + Node Exporter + Telegram alerts
  • Automated node deployment with Bash, systemd service management

Blockchain Ambassador & Node Operator — Aleo Protocol (Mar 2023 – Feb 2025)

  • Operated Aleo full node (ZK protocol)
  • Technical content creation, community onboarding, node setup guides

Projects

claude-multi-agent-assistant Multi-agent personal assistant on Claude Code: thin router + 7 specialized subagents, file-based memory in git, deterministic fail-closed guardrail hooks (PreToolUse data isolation, PostToolUse validation), mentor-mode agent with a 4-level hint ladder. Terraform for AWS deploy. MIT.

node-infra Bash-based node monitoring and deployment toolkit. Includes systemd service management, Docker containerization, Nginx reverse proxy, and GitHub Actions CI pipeline with shellcheck.

monitoring Production-ready monitoring stack: Prometheus + Grafana + Node Exporter with Telegram alerting and persistent storage via Docker volumes.

devops-board Offline-first Kanban task manager as a single HTML file. Drag & drop, subtasks, timers, deadlines, reminders, Telegram integration, EN/RU/UK multilingual support. No dependencies, no server required.

celestia-infra Terraform infrastructure for Celestia light node + monitoring stack on AWS. Two-server deployment: node + Prometheus/Grafana. One-command deploy and destroy.

ai-server-assist AI-powered server management agent using AWS Bedrock + FastAPI. Supports Claude, Llama, Nova, Mistral with shell execution via web UI.

celestia-node Automated deployment script for Celestia light node on Mocha testnet. One-command install with systemd service setup.

arc-node Deployment automation for Arc Network node (EVM-compatible). Dual-service architecture: Execution Layer + Consensus Layer via systemd.


Currently Learning

LLM serving (vLLM, Ollama) · RAG (Qdrant, pgvector) · LLM observability & evals (Langfuse) · Kubernetes · Python automation


Contact

LinkedIn · Telegram

Pinned Loading

  1. ai_server_assist_aws ai_server_assist_aws Public

    AI server-management agent: AWS Bedrock (Claude, Llama, Nova, Mistral) + FastAPI with shell execution via web UI

    HTML

  2. node-infra node-infra Public

    Bash toolkit for node deployment and monitoring: systemd services, Docker, Nginx reverse proxy, GitHub Actions CI with shellcheck

    Shell 1

  3. x_grok_tg_bot x_grok_tg_bot Public

    Telegram ↔ Grok relay: Python bot that forwards Telegram messages into a Grok session on X and parses replies back to the chat

    Python

  4. claude-multi-agent-assistant claude-multi-agent-assistant Public

    Personal multi-agent assistant on Claude Code: router + subagents, file-based memory in git, deterministic guardrail hooks, mentor-mode agent. Terraform for AWS deploy.

    HCL

  5. devops-board devops-board Public

    Offline-first Kanban in a single HTML file: drag and drop, subtasks, timers, deadlines, Telegram integration, EN/RU/UK

    HTML

  6. celestia-infra celestia-infra Public

    Terraform IaC: Celestia light node + Prometheus/Grafana monitoring on AWS. One-command deploy and destroy

    HCL