As a self-taught developer, my primary means of learning has always been through breakable toys—that is, empty canvases that welcome failure, where the only true goal is learning itself.
Some projects, like SubjectMatter, are an attempt to grok machine learning concepts through the only language I knew at the time, Ruby (I've since learned that a language with stronger ML support, like Python, would be better suited to the task). Others, like Loupe, are purely academic pursuits to push WordPress to its limits. Tombstone is probably the most practical (and complete) of the bunch, being a Ruby on Rails-backed simple state machine to manage projects for my team at Communicorp—complete with months-long running job queues using Redis, email notifications with Mandrill, and admin functionality.
Though these aren't complete projects persay, they hold a special place in my little dev ❤️.
Personal blog and portfolio written in Jekyll and hosted on GitHub Pages.
System-specific files for getting a new computer environment up to speed.
Manage projects with approval processes, email notifications and a dashboard.
Determine if a marketing email subject line is spammy with machine learning.