Experience
Agentic AI
I build AI systems security-first — grounded in real-world experience deploying multi-instance systems, not theoretical frameworks. Here's where I focus.
AI Assistant Implementation
Persistent, context-aware assistants with multi-instance architecture, long-term memory, tool integration, and human-oversight controls.
Agentic DevSecOps
Security-operations AI handling alert triage, incident response, audits, and pipeline integration — with documented reasoning and reversibility.
AI Security Guardrails
Prompt-injection defense, output validation, tool-call governance, and permission boundaries for agentic systems.
Vibe Coding Enablement
Helping teams adopt AI coding tools — with code-review practices and collaboration habits that keep quality high.
On GitHub
The projects behind the work
The agentic-AI systems I run are public. These are the actual codebases — the persistent, multi-instance assistant framework, and the security layer that keeps it inside its guardrails.
Building something agentic?
Whether it's a production assistant or guardrails for a system you already run, I'm always glad to compare notes.
Get in touch