About

I'm Richard Callis, an engineer working under the Khalisio name. I focus on the gap between LLM demos and production systems — the part where you have to deal with retries, cost ceilings, observability, and the unglamorous-but-load-bearing infrastructure underneath.

Most of my recent work is on agent systems: how they call tools, how they share memory, how they coordinate, and how you actually run them on infrastructure you control. My homelab is a working proof of all of it — you can read the breakdown on the portfolio page.

I'm interested in projects where the AI piece sits inside a real business system — not a chatbot wrapper around an API, but agents that touch databases, file systems, ticketing tools, and code repos with proper guardrails. If that's what you're building, we'll probably get on well.