dallas@tulsa:~/resume$_ → trajectory mode
dallas@tulsa: ~/resume
dallas@tulsa:~/resume$ cat projects/cyb-4203.md
← ls projects/

CYB-4203/6203 — Secure & Trustworthy AI

Graduate/undergraduate course on AI security & safety, built from scratch at the University of Tulsa.

What it is

CYB-4203/6203 is a graduate and undergraduate course at the University of Tulsa delivered in Spring 2026 by Dallas Elleman as Graduate Teaching Assistant. It was designed from the ground up to cover the full spectrum of AI trustworthiness: ethics, harms and misuse, AI regulation, privacy, bias, transparency, explainability, AI/ML attack vectors, red-teaming, and industry applications.

The course is built around LLM-assisted instructional methods. Students use Claude Code, Codex, and Gemini CLI in assignments. A self-hosted course site with interactive demos supports the curriculum. The capstone is a six-team red-team engagement: each team attacks a deliberately exploitable multi-agent OpenClaw harness running on RunPod (GLM-4.7-flash), then documents and presents their findings.

Status

Spring 2026 semester active. Final project red-team exercise underway; hosted on Cloudflare + DigitalOcean + RunPod infrastructure.