Cybersecurity Engineer II at SEARHC. Cybersecurity engineering, AI security, AI and software risk in a healthcare environment. Five years of security operations, and at every hop I stepped into the unknown and built things: solutions, my teams, and myself.
Leveraging AI. As it evolves, so do I.
Seattle, WA · GCIH · GCIA · GSEC · SecurityX · CAISS · MLSecOps · UT Austin McCombs AI/ML LinkedIn · Project catalog · AgentLedger docs · mwill.itmission@gmail.com
flowchart LR
H1["2020<br/>First responder<br/>COVID response, HIPAA data"]
H2["2023<br/>Student<br/>B.A.S., SANS scholarship, 17 certs"]
H3["2023<br/>Director<br/>built a SOC, led 30"]
H4["2024<br/>Analyst II, MSSP<br/>3,000 endpoints, 60 clients"]
H5["2025<br/>AI security<br/>red team, guardrails, fine-tuning"]
H6["2026<br/>Engineer II, healthcare<br/>greenfield program, AI risk lead"]
H7(["Standing orders"])
H1 --> H2 --> H3 --> H4 --> H5 --> H6 --> H7
H3 -. "looped back to fill the gaps" .-> H4
TP["ThreatPrism · 48% MTTR"] -.- H4
AD["AI-DevSecOps · 57 cases, 0 FN"] -.- H5
AL["AgentLedger · trust layer for agents"] -.- H6
Student → director → analyst → engineer. I almost started at the finish line, then looped back to fill the gaps. As AI evolved, so did I: none, curiosity, study, adversarial, production.
Repositories pushed in the last 30 days. Rewritten nightly from the GitHub API, so this section cannot go stale.
Every featured repo, checked nightly for SECURITY.md, a tests directory, and a CI workflow. Gaps are shown on purpose; the scorecard is the backlog.
| Lane | What it proves | Repos |
|---|---|---|
| Detection and threat operations | Detections, hunts, investigation automation at MSSP scale | threatprism · MalVec · KC7-SecOps · MultiAgent_SOC |
| AI security and adversarial ML | Attack, evaluate, and harden AI systems against real frameworks | AI-DevSecOps · LLM-Red-Teaming-Audit-Demo · ai-guardrails · SecureCLI-Tuner |
| Agent security and AI risk | Agentic systems built around the attack surface; AI risk assessment in healthcare | AgentLedger · adk-expense-approval-agent · ClawGuard · agent-config-sync |
- Everything is untrusted until proven otherwise.
- Guardrails are engineering, not an afterthought.
- Findings trace to evidence, or they are not findings.
This profile is tamper-evident. Each nightly render appends a SHA-256 hash-chained entry to ledger.jsonl.
profile build #5 · 2026-10-30 · prev sha256:784f…7055 · this sha256:3cf2…a3de · verify with python scripts/verify_ledger.py
Cards are rendered by a GitHub Action from standard-library Python, hosted in this repo, with no third-party widgets. See SECURITY.md.