Infrastructure automation.

HDC server provisioning

Developed reusable Hyper-V provisioning workflows around PowerShell, manifest inputs, and cloud-init. Ubuntu 24.04 and Rocky Linux 10 provisioning was verified with static addressing, VLAN assignment, SSH, cloud-init, and DNS.

A C# interface is being developed with separate Models, ViewModels, and Services, including validation and command-state refactoring. The WPF migration, job scheduling, packaging, and later portal fixes remain development work rather than a completed product.

Container delivery and AI assistant integration

As a longtime fan of the Mass Effect series, I named my AI infrastructure assistant EDI in tribute to the games' artificial intelligence character. The project explores AI-assisted infrastructure monitoring and troubleshooting.

Migrated portfolio delivery to GitHub Actions and GHCR while retaining a private GitOps repository. Builds publish both a latest tag and a commit-specific tag; Argo CD Image Updater configuration targets image digests.

Implemented a Python FastMCP service in the repository with read tools for nodes, unhealthy pods, logs, events, Argo CD, Longhorn, and CloudNativePG. Kubernetes and Hyper-V assistant integration is ongoing; repository implementation is distinct from a verified live assistant connection.

PowerShell for enterprise administration

PowerShell is the primary automation language for Windows and Active Directory work.

• Query and export computer and user data from Active Directory.
• Normalize input files and enrich records with logon names, email, title, and department.
• Automate group membership, local-account removal, firewall configuration, domain operations, and server inventory.
• Install prerequisites for System Center products and validate server readiness.
• Collect reboot history, health information, and deployment results from remote systems.

GitOps as operational automation

The Kubernetes environment treats Git as the desired state. Argo CD reconciles applications, reports drift, retries failures, prunes removed resources, and provides a visible deployment history.

• App-of-apps repository structure.
• Kustomize-rendered application manifests.
• Automated image digest updates for selected workloads.
• Sealed Secrets for encrypted secret workflows.
• Validation, commit, push, sync, rollout, and smoke tests packaged into complete Bash scripts.

Infrastructure as code

Terraform, Ansible, and Packer are used in the home lab to explore repeatable virtual-machine provisioning, configuration, and image creation. The objective is to move from manually assembled systems toward reproducible environments with clear inputs and versioned changes.

Design principles

• Safe defaults: scripts stop on errors, validate prerequisites, and avoid overwriting unknown state.
• Idempotence: repeated execution should converge or exit cleanly.
• Complete deliverables: full scripts and configuration files reduce copy-and-paste mistakes.
• Evidence: every automation ends with verification output rather than assuming success.
• Rollback: Git history and declarative state make changes reviewable and reversible.