██████╗ ██╗ ██╗████████╗███████╗██████╗ ██╗ ██╗███╗ ██╗███╗ ██╗██╗ ██╗
██╔══██╗╚██╗ ██╔╝╚══██╔══╝██╔════╝██╔══██╗██║ ██║████╗ ██║████╗ ██║╚██╗ ██╔╝
██████╔╝ ╚████╔╝ ██║ █████╗ ██████╔╝██║ ██║██╔██╗ ██║██╔██╗ ██║ ╚████╔╝
██╔══██╗ ╚██╔╝ ██║ ██╔══╝ ██╔══██╗██║ ██║██║╚██╗██║██║╚██╗██║ ╚██╔╝
██████╔╝ ██║ ██║ ███████╗██████╔╝╚██████╔╝██║ ╚████║██║ ╚████║ ██║
╚═════╝ ╚═╝ ╚═╝ ╚══════╝╚═════╝ ╚═════╝ ╚═╝ ╚═══╝╚═╝ ╚═══╝ ╚═╝
A researcher and tinkerer somewhere between machine learning theory and bare-metal code. I write Python when I think fast, C when I think carefully, and stare at kernel logs when I think too much.
Currently exploring cognitive architectures and what it means for a system to know something — not simulate knowing, actually hold a state that persists.
"Most systems react. The interesting ones remember why."
🔭 AI research · ⚡ always compiling something · 🐇 fast, occasionally reckless
- Whether "understanding" needs embodiment, or just enough recursive structure to fake it convincingly
- Active inference vs. backprop-trained policies — different roads, same destination?
- Where exactly a system stops predicting and starts deciding
My friend Stell is building Anima — an experimental cognitive architecture exploring active inference, IIT, and decision-making without leaning on an LLM as the reasoning core.
If that sounds like your kind of rabbit hole, a ⭐ star, 👀 watch, or 🍴 fork would go a long way — and I'll happily return the favor on your projects too.
Open to interesting problems — from quick fixes to deep architecture work. No task too complex, no rabbit hole too deep. If you've got something worth building, let's talk.
