Figuring Things Out
Background, DEV Environment, Goals, and Writing
This is the first blog I’m writing, so “figuring things out” is an understatement. Let’s start with who I am and how I got here. My name is Chris Kornaros, I’m an Analytics Engineer with a BSM and MS from Tulane University. I’ve always been passionate about learning. During my masters program in business analytics, I became obsessed with data analytics and the science behind it— especially because of how fast the field grows and changes. Since then, my learning and research interests are primarily around data engineering, architecture, optimization, and, more recently, applying that to AI. To enhance my learning, I bought a basic MacBook Air and a Raspberry Pi 4 in 2024. I’ve been using those to learn and build on, with minimal configurations for both (the Mac has now grown a bit, more on that in another post).
I’ve been using LLMs / AI since about 2023, and I’ve been paying for a Claude Pro subscription since the end of 2024. Originally, I used Claude because I preferred the project layout, response styles, etc. to ChatGPT. To start, I was basically using Claude as a tutor, helping me learn my way around Linux environments and building a personal website. It was great, but I never really got the hype, even as it felt like everyone in my field was deploying agents and doing X, Y, or Z with RAG, then with OpenClaw to run autonomously, and so on.
About a month ago, I decided to take full advantage of my Claude Pro subscription. While I hadn’t done anything advanced yet, I was familiar with the horror stories of burning through usage and the newly minted ~tokenmaxxing~. The first thing I did with Claude Code (I’m using the VS Code extension) was setup a private ‘agentic_optimization_research’ repository. Learning is great, but spending tons of money to do so does not always feel fun. The general goal was (is) simple, how do Claude and I work together more efficiently based on the limits of a Pro plan?
In the last month, I’ve learned a lot and I’m really enjoying the freedom Claude Code provides. There are some basic things I’ve built to test out the research I’ve done— I’ll do separate blog posts for those, once the apps are in a better spot. As for what I’ve learned in my research repo so far, well it’s taking a while to write that up.
For now, stay on the look out for mini-guides and blogs that cover some of my workflows. For example, everything I’m doing is in an Obsidian vault, which is live synced across all my machines. Setting this up just takes a little work, but the benefits are incredible, I’ll be sharing that soon. On what you can expect, I try to keep blogs between 500 and 1,500 words, my longer ‘guide’ blogs may get to 3,000, but that’s it. For anything far longer or more in depth, that lives directly on my website, under the guides section (my raspberry pi server guide is 30,000 words). So, the options are there for those who want really in depth technical walkthroughs, and those who want something that is more to the point.