01
AstroCloudflare PagesCSPAEO
Migrating an AEO Agency Off Squarespace
An agency selling AI-search readiness can't run on a platform that hides the technical layer that readiness depends on. Rebuilt Penpixel Creative's own site from Squarespace into a custom Astro build on Cloudflare Pages — scoped CSP, HSTS with preload, Cloudflare Turnstile on the contact form, and a robots.txt implementing the Content Signals Policy standard with explicit AI crawler allowlisting. The site is the proof of the service.
Astro · Cloudflare Pages · CSP + HSTS preload · AI crawler allowlisting
02
TerraformAWS EC2Auto ScalingALB
28 AWS Resources for $22/Month
Built a full auto-scaling AWS infrastructure with Terraform — EC2, ALB, ASG, security groups, and more. Documented the 5 things that went wrong along the way so you don't have to find them yourself.
28 resources · 5 failures documented · $22/month
03
TerraformAWS EC2JenkinsAL2023
The Jenkins Job That Took 6 Tries
Provisioned a Jenkins CI/CD server on EC2 with Terraform. Hit six distinct failure points — hardcoded AMIs, Java version mismatches, AL2023's switch from yum to dnf, GPG key rotation — and documented every one.
6 failures → 1 working pipeline
04
Docker SwarmAWS EC2WordPress
Self-Healing WordPress Production Stack
Three-node Docker Swarm cluster on EC2 running a production WordPress stack. When a container dies at 2am, the orchestrator brings it back. No pager, no humans required.
WordPress + MySQL · 3-node cluster · Service auto-recovery
05
AWS PollyGitHub ActionsS3
Multilingual TTS Pipeline
AWS Polly neural text-to-speech pipeline with full GitHub Actions CI/CD. Converts text to audio across multiple languages on every push — OIDC auth throughout, zero stored AWS keys.
Neural engine · OIDC auth · Zero stored keys
06
Docker SwarmAWS EC2Overlay Networking
Docker Swarm Self-Healing Container Infrastructure
Three-node Docker Swarm cluster on AWS EC2 with global service deployment and overlay networking. If a node or container fails, Swarm detects and recovers without human intervention. Built to replicate real production failure conditions.
1,555 LinkedIn impressions · Benchmark post
07
BashLinuxsystemd
Linux Security Audit Script
Automated security audit with systemd timer. First run flagged Apache2 exposed and Open WebUI bound to 0.0.0.0. Reduced flagged services from 3 to 0. Snap storage dropped from 54G to 30G — a 44% reduction.
3 flagged services → 0 · 44% storage reduction
08
BashLinuxsystemd
Linux Process Sentinel
Bash service watchdog with autonomous recovery. Monitors critical processes, detects failures, and restarts services without human intervention. Built and tested against live Ubuntu infrastructure.
Autonomous recovery · No pager required
