I built
https://learnfrontend.com as an effort for me to A) teach people front-end development and B) to quit my consultant job so I can focus on making a difference in people's lives. The site features a news aggregate, videos, articles, goals, and interactive training platform.
The site is built on Symfony 2.5, PHP 5.5.x, Postgres, AngularJS 1.2.x, and SCSS. I use Jenkins for CI, Doctrine2 as the ORM, Redis for session I/O, Composer/Bower for dependency resolution, GitHub for documents, tickets, code reviews (pull requests to my self). HAProxy is on standby for the need to load balance and I have a job on Jenkins to scale up and down to a box if needed. I'm currently adding a Docker/Vagrant setup to I can easily manage dev environments (I don't think I'll do the same for production, we'll see). Stripe handles payments, I bought some cheap SSL cert (Stripe handles PCI info outside the server I/O). New Relic is used to monitor both QA, production and my jenkins deploy box. DNS is handled by GoDaddy. I have a Cron handling the news aggregate refresh, diff logic to populate new articles to a mobile curator app I built, and an approve/deny system that determines which articles are allowed to be propagated to the users. Videos are hosted through a third party service and assets internally are handled stored through S3 with CloudFront URL signing assets.
Although it just released, I have received a large amount of important feedback thats going to dramatically change the site. This includes offering more to the user for free, create more interactive components (my writing is too boring / too technical), and improve the marketing (showing the actual tools). I have some POCs in the works for a simulated IDE on edge Chrome/Firefox (uses WebWorkers to observe stackoverflow-able code w/o crashing browser, ScalaJS and Angular port on ScalaJS as MVC frontend; the experience is very much a level up from CodeAcademy's IDE-tutorial).
I'm also in a transitionary period where I want to build the new tools and reconstruct the existing ones in Scala (using Play as the framework). I have more time than before since I saved a significant amount of money to pursue this (I built v1 part time, 30 hours a week for 6 weeks with 30k lines of code added). The force behind this decision is I want more experience with functional programming while leveraging what I know already (JVM / Java).
One of the critical experiences I've had was actually materializing an idea and shipping it. Failing fast and often is an important strategy that's working for me after launching v1.
TL;DR: Built an e-learning site. I learned from failures. Building next version thats going to be rainbows and shit (thats the plan at least).