Security × AI Automation

Devin Anderson
building & securing with AI.

Cybersecurity and GRC professional who builds and secures AI automation workflows using AI agents and LLMs. Self-taught, hands-on, and open to full-time or contract roles.

Security meets automation.

I come from a IT, cybersecurity and GRC background. Risk assessments, compliance frameworks, third-party risk, and policy work. But as AI started reshaping how teams operate, I saw a gap: most organizations are adopting AI workflows and automation faster than they can secure or govern them.

So I started getting my hands dirty. Instead of just reading about AI security, I'm building real projects. Triage automation, observable AI workflows, and systems that plug into the tools teams actually use. Every project I work on is something I can walk through, explain, and defend.

My value proposition is simple: I sit at the intersection of security knowledge and technical building. I understand the compliance and risk side, and I can build the AI-powered automation that makes teams faster, more consistent, and more observable. That's the gap I fill.

Tools & Technologies

Python (Intermediate) FastAPI Docker n8n Claude Langfuse Jira Slack AWS REST APIs Git / GitHub Zapier OneTrust Confluence Hubspot Microsoft Office Suite ChatGPT Gemini Vibe Coding

What I've built.

Real systems, not tutorials. Each project is designed to solve an actual operational problem with production-grade tooling.

Module 2 · Complete

Red-Team & Regression Testing

Red-teamed the triage logic with 9 adversarial test cases: contradictory signals, missing data, and prompt injection attempts. Replaced rule-based keyword matching with a controlled LLM (Claude API), built production-level guardrails, and created an automated regression suite. Result: 3/9 to 8/9 passing.

Claude API Python FastAPI Docker n8n Langfuse
9 adversarial red-team test cases
Prompt injection defense layer
LLM-powered triage with system prompt
Input validation & sanitization
Rate limiting (30 req/min)
Automated regression test suite
Structured scoring & pass/fail reporting
3/9 to 8/9 documented improvement
Module 1 · Complete

AI Security Triage Automation

End-to-end security event triage: intake, normalize, AI-assisted classification, Jira ticketing, Slack alerts, and Langfuse observability. Every run is traced, ticketed, and auditable.

n8n Python FastAPI Docker Jira Slack Langfuse
Webhook intake with auth
Field normalization layer
AI triage with rule-based fallback
Structured Jira ticket creation
Real-time Slack SOC alerts
Full LLM observability via Langfuse
Containerized microservices
Artifact storage per run

Coming next

Module 3 · Planned

AWS Findings Automation

Security Hub and GuardDuty integration, S3 evidence storage, and automated cloud security finding triage.

Project A · Planned

AI Triage Red Team Benchmark Suite

OWASP LLM Top 10 adversarial evaluation framework with Promptfoo, Garak, and GitHub Actions CI/CD.

Let's connect.

I'm open to full-time and contract roles in Cybersecurity, GRC, and AI automation. If you need someone who can bridge compliance and technical implementation; or if what I've built here resonates; I'd love to connect.