Software that works.Systems that scale.

Problem

Every new application starts with the same setup (structure, authentication, pipeline, infrastructure) before any feature exists.

Approach

Switchable authentication (JWT, Microsoft Entra ID, or none), optional UI, tests, samples, Docker, and infrastructure, keyless OIDC deployments from GitHub Actions to AWS, and Terraform for dev and prod. A verify script generates, builds, and tests every flag combination before each release.

This is an open-source, MIT-licensed version of the scaffolding I built for my team, rebuilt from scratch for public use.

Problem

Scheduled data jobs often start as one-off scripts: credentials in source, errors silently ignored, SQL built by string concatenation, and no tests. They run, but they can’t be handed to a team.

Approach

Secrets resolved from environment variables or a SecretManagement vault, REST calls with timeouts and exponential-backoff retries, and parameterized, batched inserts that roll back as a unit. Table and column names are validated before a connection opens, and -WhatIf previews a run without touching the database. Pester tests and PSScriptAnalyzer run in CI on all three operating systems.

Problem

Tail Blazer was built on WPF, which limited it to Windows.

Approach

Adapted it to Avalonia UI and .NET 10 while keeping its reactive architecture, so log inspection stays fast and familiar.