After 15 years of building Rails apps, the AI wave presented an interesting challenge: how to integrate these new tools without abandoning everything that already works. When ChatGPT launched in late 2022, it became clear that ignoring this shift wasn’t an option. The problem with most AI tutorials is they assume you’re starting fresh or want to rebuild everything from scratch. That approach doesn’t make sense when you have years of Rails experience and existing applications to maintain. The solution has been gradual experimentation. Some approaches work well, others don’t. This site documents that ongoing journey, both as a reference and hopefully something useful for other developers facing similar questions.

​
What’s Here

Articles - A few guides so far covering tmux workflows and connecting Rails apps to Python AI services. More posts planned as new patterns emerge. About - Additional background and current projects.

​
Current Learning Areas

Rather than claiming expertise, this is an honest look at what’s being figured out in real time. The focus is on learning Python alongside Ruby, understanding how LLMs actually function, and finding practical ways to integrate AI into existing Rails applications. Current work involves several independent SaaS projects, testing where AI adds genuine value versus where it’s just noise. The approach is methodical but acknowledges there’s plenty of trial and error involved. Key lessons discovered so far:
  • Rails and Python can work together more smoothly than expected
  • Docker simplifies managing multiple services significantly
  • AI APIs fail frequently, making robust error handling essential
  • Most integrations are actually simpler than tutorials suggest

​
Background

My journey started in 2006 working with various startups, adapting to whatever role was needed - frontend, backend, fullstack development. Six years were spent at SurveyMonkey building Rails applications for survey tools and data processing systems. Before and since, the work has involved Rails consulting and independent projects. When AI became mainstream, the choice was learn or risk obsolescence. Rather than abandon 15 years of Rails knowledge, my goal became bridging these two worlds effectively. My current focus involves AI-powered SaaS development while documenting useful patterns. No grand strategy - just sharing what works and what doesn’t as it gets figured out. Contact: