Hey, I’m RC.
Co-founder and CTO of Summation, an AI-native platform for business planning and decision-making.
I build LLM-powered products at the application layer, the messy, rewarding part where a model meets real users, real latency budgets, and real cost constraints. Lately I’m deep into coding agents: Claude Code, Codex, and whatever sharp open source project shows up next.
Why this site
The name comes from a belief that learning has a loss function: a measurable gap between what you know and what you need to know. Every miss carries signal. Every failure leaves a trail, missing context, leaky abstractions, bad defaults, weak verification.
A good harness turns an LLM into a dependable coworker. A bad one turns it into a slot machine. I write here to figure out the difference. Start here.
How I write
I follow the Feynman rule. If I can’t explain it clearly, I don’t publish it. If I can’t make it visual, I probably don’t understand it yet. Expect illustrations, working code, and examples with real numbers. No hand-waving.