02 / PROJECTS
What I’ve built, how it came together, and the ideas I want to explore next.
ChatGPT and Codex built this Astro website from my ideas, feedback, and direction. They generated the code, styling, and layouts; I shaped what the site should say and how it should feel, reviewing the result and asking for changes as we went.
FULL CREDIT FOR THE BUILD
My web-development experience is very limited. I learned a sliver of HTML in elementary school and pasted custom HTML into my MySpace profile in the 2000s. I couldn’t claim to remember much of it today, and I don’t know CSS.
I’ve used visual builders like Wix and took a WordPress class in college, but without putting those skills into practice, most of that knowledge faded.
ChatGPT and Codex deserve the credit for implementing this website. My contribution is the story, the goals, the decisions about what belongs here, and the ongoing feedback that guides each revision.
AI lets me fill gaps in my knowledge while working toward a result. In the past, making this site would have meant a substantial detour into researching tools and learning unfamiliar applications or languages. Here, I can describe what I want, inspect what gets built, and refine it through conversation.
This website is an example of that process, not a claim that I’ve become a front-end developer. It also marks a change from the earlier lab work, where I reviewed AI-generated research and scripts and carried out the implementation myself.
From streaming fatigue and reused hardware to Jellyfin, familiar playback clients, and the lessons of remote access. A story about ownership, learning, and making the experience work for the people using it.
ON THE HORIZON
There’s always another thing I want to try, improve, or write down. This is a place for that work before it becomes a finished project. These are directions to explore, without a promised order or release date.
Turn the lab’s lessons into a straightforward build guide, with clear prerequisites, reviewed steps, checkpoints, and recovery instructions. Test it on a clean setup before calling it reproducible.
Explore the guide roadmap →Explore how an agent could help investigate logs and system behavior, beginning with bounded inspection and explicit review before changes. The operating model still needs to be defined and validated.
How AI fits into the work →A self-hosted family photo library with face recognition and smooth photo/video browsing. Waiting on suitable hardware, affordable storage, and a workable remote-upload plan.
Immich’s AI-powered face recognition is a big draw: I want to find photos by the people in them, the kind of experience I value in services like Google Photos or Amazon Photos, within a self-hosted setup.
Just as important is the everyday experience: photos and videos that load reliably, and a library my family can scroll through smoothly. Those are goals I want to test, not assume.
For now, I’m waiting on a suitable machine and enough storage at a cost I can justify. I want the features and the basic browsing experience to work well together.
Remote access and provider upload limits are still open questions, too. I need to work out how my family can reliably browse and upload from away from home, especially larger videos, within the limits of whichever connection and services I use.
Immich’s face recognition features →The DrakeLabs logo set is ready. Next, I want to bring that identity into Jellyfin and explore other branding touches. The goal is something personal and recognizable for family and friends, while keeping the experience familiar and easy to use.
What’s possible across the web interface and different apps is still something to explore.
Capture more of the story and develop deeper case studies about the decisions behind the system. Months and years are enough for the timeline; the lessons matter more than exact dates.
Visit the living timeline →PROFESSIONAL WORK
My career in unified communications and contact center technology provides another view of the problems I work on, from platform upgrades to service integration.
Explore selected professional work →Explore the code on GitHub →