
healthcare · data · ai
Chatting about AI
And keeping it real.
This website is a collection of my experience and thoughts while building with AI. My primary focus is educational, where I attempt to simplify the key concepts of “Applied Agents” and share ideas around the LLM application space. I feel somewhat confident speaking on the topic because I am currently spending way too much time behind the computer, creating agentic applications through the help of agents.
In addition, I’m building a career in this space. This isn’t my formal business page, but let’s connect. I love talking about LLMs, and most importantly how we can maximize their use to spend less time behind a screen and more time in the real world.
A couple of notes:
- AI slop is ruining the internet (blog post coming later). That said, yes, I will be working with agents to help build out my content. My promise is to prefer quality over quantity, and ensure the content is at least well curated.
- On the design — I spend a lot of my time in Claude Code terminals to manage all aspects of my work. Terminals are funny. Folder structure matters a lot. I’m trying to incorporate some of that “feel” throughout this website.
featured
How AI Agents Actually Work
Break down the 5 layers of an AI agent — from the model to tools to memory. Built for curious people, not engineers.
start the guide →
interactive labAI Conversation Simulator
Watch a real AI API call happen step by step. See the system prompt, the messages array, and how the model responds.
open simulator →
essay90M Records, One Eval Script
Karpathy's autoresearch pattern applied to Medicare fraud detection. An AI agent iterated overnight on 90M CMS claims records — AUC 0.556 to 0.810 in 18 iterations.
read essay →