Sebastian Malieni

sevakunjadas.bms@gmail.com Buenos Aires, Argentina GitHub LinkedIn

Summary

Junior DevOps Engineer with hands-on experience building reproducible Fedora-based environments using Ansible, QEMU/KVM, and cloud-init. Currently developing a Hardened Edge-Gateway & Containerized (Docker/K3s) Infrastructure under the mentorship of a Senior Cybersecurity Architect, with a core emphasis on security, Infrastructure Observability (GLP Stack/Uptime Kuma), and 'Zero-Trust' principles. Leverages AI-assisted development (LLMs) to bridge technical gaps and accelerate the delivery of complex automation workflows efficiently.

Skills

Infrastructure & Virtualization: Fedora Server (RHEL ecosystem), QEMU/KVM, Libvirt, Cloud-init, AWS (Solutions Architect path).

Networking & Security (DevSecOps): MikroTik RouterOS, WireGuard VPN (Hub & Spoke), VLAN Segmentation (802.1Q), Firewall/NAT hardening, Zero-Trust network isolation, Pi-hole (DNS Security).

Orchestration & Containers: K3s (Kubernetes), Docker & Docker-Compose, Nginx Reverse Proxy, Ingress Management.

Automation & Methods: Ansible (Playbook development), GitHub Actions (CI/CD pipelines), Terraform, AI-Assisted DevOps (LLM Prompt Engineering), Git, Shell Scripting (Bash/Fish).

Observability & Monitoring: Grafana, Loki, Prometheus (GLP Stack), Uptime Kuma, Alloy (Log shipping), Node Exporter.

Project Experience

Guided by a Senior Cybersecurity Architect.

  • Secure networking: Engineered a Hub-and-Spoke VPN architecture using MikroTik and WireGuard to securely integrate remote appliances (IoT, smart TVs, admin clients) into a private, VLAN-segmented environment.
  • Host virtualization: Deployed and managed a Fedora-based KVM/QEMU virtualization host, utilizing bridge networking and custom systemd sequencing to ensure high availability of core services.
  • Container orchestration: Orchestrated a K3s (Kubernetes) cluster to host self-healing microservices, managing persistent storage (PVCs) and traffic routing via Nginx Ingress.
  • Security & isolation: Implemented a Zero-Trust model by enforcing strict inter-VLAN firewall rules on RouterOS and centralizing network-wide DNS security via Pi-hole.
  • Automation as code: Developed Ansible playbooks for automated VM provisioning and environment consistency, leveraging AI-assisted development to rapidly iterate on complex automation logic.
  • Infrastructure observability: Designed and deployed a centralized monitoring platform using a GLP stack (Grafana, Loki, Prometheus) and Uptime Kuma. Implemented proactive alerting and real-time dashboarding to track host metrics, container health, and network uptime across the hybrid environment.
  • Management isolation: Architected the observability stack within a dedicated QEMU VM to ensure monitoring persistence during cluster-wide maintenance or failures.

Production infrastructure delivered to and actively running at Fundación Huésped — a Buenos Aires HIV/AIDS research NGO that tracks biological samples across freezers and liquid nitrogen storage.

  • Container packaging: Containerized a legacy PHP/Apache biobank application using Docker Compose, enabling reproducible, single-command deployment on any Linux machine — replacing a fragmented manual setup process the foundation adopted immediately into production.
  • Service orchestration: Designed a multi-service Docker Compose stack (web, database, backup) with strict network isolation, placing the MariaDB instance on an internal-only Docker network unreachable from outside the container stack.
  • Security & isolation: Applied least-privilege DB credentials, enforced a no-secrets-in-repo policy via .env templating, and architected network segmentation to minimize the attack surface of an NGO production system.
  • Data resilience: Implemented an automated daily mysqldump backup service with 7-day rotating retention, ensuring operational continuity without exposing credentials in the process list.
  • Cloud architecture design: Proposed a cost-optimized AWS migration (EC2 + Nginx + Terraform) tailored to NGO budget constraints, delivering infrastructure-as-code for the proposed cloud environment.
  • CI pipeline: Integrated GitHub Actions to validate Docker builds and Terraform configurations on every push, ensuring infrastructure changes are tested before reaching the production server.

GitHub Actions CI/CD pipeline deploying a SvelteKit site to GitHub Pages via a two-repo GitOps pattern.

  • Pipeline design: implemented a GitHub Actions workflow triggered on every push to main, running a sequential quality gate (type-check → build) before deploying — ensuring broken output never ships.
  • Two-repo GitOps: source code stays in a private repository; only the compiled artifact is pushed to a separate public repository served via GitHub Pages at a custom domain.
  • Reproducible builds: replaced npm install with npm ci to guarantee lockfile-exact installs across all runs, eliminating dependency drift between local and CI environments.
  • Least-privilege deployment: scoped the Personal Access Token to the public Pages repository only, limiting the blast radius of a potential credential leak.
  • Dependency caching: enabled npm cache on the Node.js setup step to reduce install time on successive runs.
  • Pinned action versions: all third-party actions locked to major version tags to prevent silent upstream changes and supply chain risk.
  • Built local high-availability lab scenarios using Fedora and QEMU/KVM virtualization.
  • Configured NGINX load balancing and cloud-init based instance bootstrapping for repeatable tests.
  • Validated fault tolerance and service distribution behaviors under controlled conditions.
  • Provisioned a RHEL 9.5 virtual lab environment with multiple disks and automated initialization via cloud-init.
  • Scripted QEMU VM lifecycle commands to streamline administration and lab repeatability.
  • Applied SSH hardening controls to improve secure remote access practices.

Certifications

  • AWS Certified Solutions Architect - Associate (in progress) — study notes.
  • HashiCorp Certified: Terraform Associate (next to do).