Rodrique Heron · self-driven walkthroughs

talks &
walkthroughs

Self-paced technical walkthroughs. Read at your own speed, pause and replay any embedded terminal session, no live presenter required.

Rodrique Heron · Red Hat SA Manager · building agentic AI on a homelab
Available
Live ~20 min read 23 slides 3 terminal replays
Agentic AI as my SRE coworker

Seven words diagnosed a homelab incident end-to-end. A Claude Code session demonstration, then the open question of building this same agentic behaviour without Claude Code as the harness — small models, local inference, the harness problem, and how Red Hat Summit 2026 just shipped the pattern I'd been prototyping.

Claude Code Agentic AI SRE vLLM RHEL AI Semantic Router MCP Local Inference
Read the walkthrough →
About
Why "walkthroughs", not "talks"
Live talks are scheduled, time-boxed, and gone. Walkthroughs are documents you read at your own pace. They have play/pause controls instead of a presenter, side-rails instead of Q&A, and the asciinema replays let you scrub through any moment that mattered.

Each walkthrough is one self-contained HTML file. No build step, no slides framework, no analytics. Open it offline, share the link, embed it — whatever you need.