Guthrie McAfee Armstrong

gmarmstrong.dev · GitHub · LinkedIn · public@gmarmstrong.dev

Site Reliability / Platform Engineer · United States · Eastern Time (ET)

Summary

Site reliability and platform engineer with four years of professional experience. Specializes in AWS, Terraform, and secure CI/CD. Active contributor to open source projects.

Professional Experience

Coalition, Inc. — Site Reliability Engineer II

Remote · Jun 2025 – Present

• Cut large Atlantis (Terraform) plans from roughly two hours to three minutes (40x) by moving execution to ephemeral storage and adding concurrency controls and plugin caching
• Deployed an organization-wide package-security proxy service and hardened shared GitHub Actions runners for npm, Python, Go, and container builds
• Contributed upstream BuildKit proxy support to route Docker build traffic through the company package-security proxy
• Led a large-scale Terraform upgrade and built reusable Go tooling for parallel planning and drift analysis across 300+ AWS infrastructure projects
• Implemented sensitive-header log redaction, verified by synthetic monitoring

New York Public Radio — DevOps/Platform Engineer

Remote · Oct 2022 – Feb 2025

• Upgraded legacy Terraform configurations to 1.x and built reusable, versioned Terraform Cloud modules
• Built standardized GitHub Actions deployment workflows for ECS, EC2, and Lambda
• Centralized workforce access to AWS accounts with IAM Identity Center, automating identity lifecycle through Entra ID, SAML, and SCIM and standardizing permission sets
• Reworked alert thresholds and context to suppress duplicate pages and speed diagnosis
• Built Python ELT pipelines and tuned PostgreSQL indexes to improve query latency

Open Source

Contributions

BuildKit (Moby) — Contributor, Proxy Networking · Go · 2026

• Designed and implemented daemon upstream-proxy support (#6996, incorporated into 71bb873)
• Fixed propagation of generated proxy-environment variables to subsequent gateway container processes (#6995, e3d3d7f, fbf541b)

Dangerzone (Freedom of the Press Foundation) — Contributor · Python · 2022

• Discovered and reported a regression that processed untrusted documents as root inside the container; the resulting fix enforced non-root execution and hardened the container against privilege escalation
• Added foundational unit tests, reduced cross-module coupling, streamlined container execution, pruned dependencies, and enabled Python 3.10 support

Projects

• dotfiles – Declarative NixOS and macOS system configurations using Nix flakes and nix-darwin
• python-datamuse (83 GitHub stars) – Python wrapper for the Datamuse API; used in creative NLP tools
• vim-muse – Vim NLP writing plugin; honorable mention at UGAHacks 2018

Skills

• Cloud & Infrastructure as Code: AWS, GCP, Terraform, Atlantis
• CI/CD & Containerization: GitHub Actions, Docker, BuildKit
• Languages & Systems: Go, Python, Bash, SQL, Linux, NixOS
• Reliability & Observability: Datadog, CloudWatch, PagerDuty
• Security & Identity: IAM, supply chain security, container hardening

Certifications

• AWS Certified Solutions Architect – Associate (SAA-C03) Mar 2024 (Exp. Mar 2027) — [click to verify]

Education

University of Georgia — B.S. Computer Science

Aug 2022

• Publication: Boots and Bail on the Ground: Assessing the Implementation of Misdemeanor Bail Reforms in Georgia, Georgia Law Review 54(4), 2020 (with A. Woods et al.) [link]
• Presentation: l-Diversity of Felony Charges in k-Anonymous Jail Data (UGA CURO Symposium, 2020)

Research Experience

UGA Rural Jails Research Hub — Intern Research Engineer

Remote · Apr 2020 – May 2021

• Built parallel Python/Selenium scrapers for 43 jail systems, keeping monthly compute costs under $20
• Parsed PDF jail records with regex/R into PostgreSQL using Java/Hibernate for cross-county analysis
• Developed containerized data pipelines on GCP with scheduling, alerting, and data-lineage tracking

Willson Center Digital Humanities Lab, UGA — Undergraduate Research Assistant

Jan 2020 – May 2020

• Linked historical census records to anonymized data on enslaved people using R and Python
• Digitized archives with Tesseract OCR and extracted structured metadata with spaCy NLP