Michael McGarrah

Designed and published a reusable Terraform module that eliminates QuickSight login friction across multi-account AWS environments. A single CloudFront distribution with a dynamically generated CloudFront Function handles hostname-based 301 redirects — zero origin servers, zero maintenance, under $0.06/month in production. Supports multiple vanity domains per deployment, ACM certificate validation, Route 53 dual-stack records, and optional access logging with a bring-your-own-bucket pattern… Show more

Designed and published a reusable Terraform module that eliminates QuickSight login friction across multi-account AWS environments. A single CloudFront distribution with a dynamically generated CloudFront Function handles hostname-based 301 redirects — zero origin servers, zero maintenance, under $0.06/month in production. Supports multiple vanity domains per deployment, ACM certificate validation, Route 53 dual-stack records, and optional access logging with a bring-your-own-bucket pattern. Published to the Terraform Registry and running in enterprise production for over a year.

code: https://github.com/mcgarrah/terraform-aws-quicksight-redirect
registry: https://registry.terraform.io/modules/mcgarrah/quicksight-redirect/aws/latest
article: Case Study - https://mcgarrah.org/quicksight-vanity-urls-serverless-edge-redirects/ Show less

My academic partner and I automated identifying the sounds of a non-English language, found the temporal boundary markers or segmentations, and fully automated output of segmentation results in a linguistically useful format. My part was generalizing the code to handle the language features and separate those into a configuration file. My partner created the non-English language dictionaries. The heavy lift was building the dictionaries for Swedish and Danish but he needed a platform to use… Show more

My academic partner and I automated identifying the sounds of a non-English language, found the temporal boundary markers or segmentations, and fully automated output of segmentation results in a linguistically useful format. My part was generalizing the code to handle the language features and separate those into a configuration file. My partner created the non-English language dictionaries. The heavy lift was building the dictionaries for Swedish and Danish but he needed a platform to use them. I also built a website to present an interface.

code: https://github.com/mcgarrah/LG-FAVE
video: https://www.youtube.com/playlist?list=PLG_DpV4CFj61_ltukKYaYQNipEFOLMqBe
demo: https://lgas.phonemes.org
pub: https://github.com/mcgarrah/LG-FAVE/blob/7c592cf6c75b8c6feee08c0fe9b79b17dd699404/Young_McGarrah_2021.pdf Show less

Extraction and visualization of point cloud data from the Microsoft Photosynth web service. Exporting the data into common formats for further analysis and processing.
Video: https://youtu.be/WZPtuNnaqVc
Code: https://github.com/mcgarrah/PyPhotoSynthExport

The goal for the final project in CS6475 AI for Robotics was to create a robotic platform to investigate computer vision technology. The platform included an Arduino with sensors and motors and a Raspberry Pi 2 for the vision and primary control system. The project URL is a video channel that shows the progress and challenges.

Develop a methodology for 10Gbps iSCSI that utilized RDMA and TOE to reduce primary CPU and memory usage. Document and managed the project as a team leader. The resulting iSER implementation was utilized in a related NetApp Phd project at the University of Bangalore India.

A web application similar to Google Maps but five years earlier and allowed for monetizing the service. Interviewed customer and defined specifications for and delivered a prototype web service for an online map delivery system. Service allowed for customers to purchase and schedule aerial imaging via aircraft or satellite imaging at varied resolutions.

Terraform automation for AWS QuickSight federated integration — converting scattered AWS blog posts and CloudFormation examples into reproducible Terraform configurations. Covers Okta SSO federation, IAM role mappings, namespace-based multi-tenant isolation, and Python Lambda layers for QuickSight API automation. Built as a reference implementation alongside the enterprise QuickSight migration at Envestnet, where the same patterns were applied to production deployments serving external… Show more

Terraform automation for AWS QuickSight federated integration — converting scattered AWS blog posts and CloudFormation examples into reproducible Terraform configurations. Covers Okta SSO federation, IAM role mappings, namespace-based multi-tenant isolation, and Python Lambda layers for QuickSight API automation. Built as a reference implementation alongside the enterprise QuickSight migration at Envestnet, where the same patterns were applied to production deployments serving external financial services clients. Companion to the QuickSight Vanity URL Redirect module (above) which handles the custom domain routing layer.

code: https://github.com/mcgarrah/quicksight_demo Show less

This application is a fully-fledged, dynamic quiz platform built on the lightweight Flask web framework for Python. The code demonstrates both front-end and back-end development using modern tools such as Bootstrap5 and dynamic rendering in the pages. Forgive me if the demo site is down… free tier hosting.

code: https://github.com/mcgarrah/legendary_quick_quiz
demo: Demo Website - https://quiz.mcgarrah.org/

A Python CLI tool that generates timesheets from git commit history — leveraging timestamps already embedded in your workflow instead of requiring manual time tracking. Supports multiple repositories, configurable work session gap detection, and output in multiple formats. Built to solve a real problem tracking hours across multi-repo freelance projects. Published to PyPI with Read the Docs documentation.

code: https://github.com/mcgarrah/git_timesheet_python
pypi:… Show more

A Python CLI tool that generates timesheets from git commit history — leveraging timestamps already embedded in your workflow instead of requiring manual time tracking. Supports multiple repositories, configurable work session gap detection, and output in multiple formats. Built to solve a real problem tracking hours across multi-repo freelance projects. Published to PyPI with Read the Docs documentation.

code: https://github.com/mcgarrah/git_timesheet_python
pypi: https://pypi.org/project/git-timesheet/
docs: https://git-timesheet.readthedocs.io/
article: Introduction - https://mcgarrah.org/generate-git-timesheets-python-module/ Show less

A Python library and CLI for importing GS1 Global Product Classification (GPC) taxonomy data into SQLite or PostgreSQL. Corrects the many-to-many relationship modeling between bricks and attribute types that other implementations get wrong — using junction tables instead of single foreign keys that silently lose data. Supports downloading the latest GPC XML from GS1, SQL export for migration, and serves as the data layer for the Nutrition API project. Published to PyPI with automated CI/CD and… Show more

A Python library and CLI for importing GS1 Global Product Classification (GPC) taxonomy data into SQLite or PostgreSQL. Corrects the many-to-many relationship modeling between bricks and attribute types that other implementations get wrong — using junction tables instead of single foreign keys that silently lose data. Supports downloading the latest GPC XML from GS1, SQL export for migration, and serves as the data layer for the Nutrition API project. Published to PyPI with automated CI/CD and Read the Docs.

code: https://github.com/mcgarrah/gs1_gpc_python
pypi: https://pypi.org/project/gs1-gpc/
docs: https://gs1-gpc.readthedocs.io/ Show less

Designed, built, and published a Jekyll plugin as a Ruby gem that automatically generates PDF and DOCX exports from Jekyll pages using Pandoc. The plugin hooks into Jekyll’s build lifecycle, supports incremental builds, Unicode cleanup, configurable Pandoc options, and works across Jekyll 3.x and 4.x. Built a complete release automation pipeline — GitHub Actions CI/CD, RubyGems trusted publishing, Read the Docs integration, and a single-command release script replacing 13 manual steps… Show more

Designed, built, and published a Jekyll plugin as a Ruby gem that automatically generates PDF and DOCX exports from Jekyll pages using Pandoc. The plugin hooks into Jekyll’s build lifecycle, supports incremental builds, Unicode cleanup, configurable Pandoc options, and works across Jekyll 3.x and 4.x. Built a complete release automation pipeline — GitHub Actions CI/CD, RubyGems trusted publishing, Read the Docs integration, and a single-command release script replacing 13 manual steps. Integrated into my resume site to generate downloadable PDF and Word versions on every build. Three-part blog series documents the infrastructure, plugin architecture, and real-world integration bugs.

code: https://github.com/mcgarrah/jekyll-pandoc-exports
gem: https://rubygems.org/gems/jekyll-pandoc-exports
docs: https://jekyll-pandoc-exports.readthedocs.io/en/latest/
article: Part 1: Release Automation - https://mcgarrah.org/ruby-gem-release-automation/
article: Part 2: Plugin Architecture - https://mcgarrah.org/jekyll-pandoc-exports-plugin/
article: Part 3: Integration & Debugging - https://mcgarrah.org/jekyll-pandoc-exports-resume-integration/ Show less

Built a Python client library from scratch for the 1WorldSync Content1 REST API when the only vendor-provided tooling was a Java SDK and PDF documentation. Provides a Pythonic interface and CLI for searching and fetching product content data — specifications, images, marketing copy, and logistics information. Published to PyPI with GitHub Actions CI/CD, trusted publishing, and Read the Docs integration.

code: https://github.com/mcgarrah/oneworldsync_python
pypi:… Show more

Built a Python client library from scratch for the 1WorldSync Content1 REST API when the only vendor-provided tooling was a Java SDK and PDF documentation. Provides a Pythonic interface and CLI for searching and fetching product content data — specifications, images, marketing copy, and logistics information. Published to PyPI with GitHub Actions CI/CD, trusted publishing, and Read the Docs integration.

code: https://github.com/mcgarrah/oneworldsync_python
pypi: https://pypi.org/project/oneworldsync/
docs: https://oneworldsync-python.readthedocs.io/
article: Introduction - https://mcgarrah.org/oneworldsync-python-module/ Show less

A comprehensive Python client for the USDA Food Data Central API with object-oriented data models, nutrient analysis with dietary reference intake comparisons, recipe analysis with ingredient parsing, and Django integration for local database caching. Includes a CLI for quick data access and an analysis tool for nutrient comparison and HTML report generation. Published to PyPI with Read the Docs documentation.

code: https://github.com/mcgarrah/usda_fdc_python
pypi:… Show more

A comprehensive Python client for the USDA Food Data Central API with object-oriented data models, nutrient analysis with dietary reference intake comparisons, recipe analysis with ingredient parsing, and Django integration for local database caching. Includes a CLI for quick data access and an analysis tool for nutrient comparison and HTML report generation. Published to PyPI with Read the Docs documentation.

code: https://github.com/mcgarrah/usda_fdc_python
pypi: https://pypi.org/project/usda-fdc/
docs: https://usda-fdc.readthedocs.io/ Show less