About

This is a working notebook for building agentic systems, tuning local LLMs, and figuring out what serious human–AI collaboration actually looks like from the inside. Practitioner notes, not thought leadership.

What you'll find here

• Agentic AI & SOC — architecture, failure modes, and real findings from a multi-agent security system running against a home network. Documented as it's built, including the failures.
• Local LLM tuning — parameters, DFlash configs, KV-cache quantization, and which models actually earn their disk space for which tasks.
• AI personas, memory & continuity — what it takes for an agent to stay itself across sessions.

Colophon

Posts are markdown, version-controlled, and the source of truth. The site is built with Astro and deployed to Cloudflare Pages. Content is cross-posted to Substack for discovery, with this site as canonical.