I build agentic AI systems
and secure the cloud they run on.
AI & Cloud Solutions Architect with 9+ years securing enterprise systems across GCP, AWS, and Azure — now architecting always-on, multi-agent AI with zero-trust guardrails.
Three ways I help
Agentic AI
Always-on multi-agent systems with capability-scoped guardrails, zero-trust isolation, and multi-model orchestration governed under NIST AI RMF.
- Multi-agent orchestration (Gemini, Anthropic, OpenAI)
- PII-egress controls & prompt-injection defense
- AI governance and deterministic guardrails
Cloud Security
Zero-trust architecture across GCP, AWS, and Azure — RBAC, WAF, encryption, and NIST 800-53 / FedRAMP-aligned defense in depth.
- Zero-trust networks & least-privilege IAM
- Cloud Armor WAF + OWASP Top 10 controls
- AES-256 at rest, TLS 1.3 in transit
Private Infrastructure
Self-hosted, sovereign deployments with egress control and private inference — sensitive data and model traffic never leave your environment.
- Self-hosted LLM inference pipelines
- VPC firewalls, Cloud NAT, egress control
- On-prem and hybrid sovereign hosting
Case studies
Multi-Site Zero-Trust SOC with AI-Driven Triage
A self-hosted SIEM/SOC, zero-trust WireGuard overlay, and per-endpoint EDR protecting multiple home networks — with an AI analyst handling first-line alert triage.
Read case study →AgentGuard MCP
MCP security gateway that sits between AI agents and their tools, scanning every call inline so PII and secrets never leave your systems — ensuring 99.9% data protection.
Read case study →Always-On Multi-Agent AI Infrastructure
A persistent agentic-AI platform where multiple specialized agents run continuously, orchestrate across LLM providers, and act on real systems — engineered for uptime, isolation, and governance from the ground up.
Read case study →Latest research
Removable Safety: How Frontier LLMs Are Stripped of Their Guardrails and Run on Workstation Hardware
A security analysis of model abliteration and 4-bit compression: the safety alignment of a ~750B-parameter flagship model can be surgically removed, compressed to 4 bits, and run at conversational speed on purchasable workstation GPUs — with the safety removal fully intact.
Seven Security Pitfalls of Always-On Agentic AI Agents
A field report from red-teaming a production always-on agentic-AI stack: seven structural security weaknesses that recur across hand-built agent systems, and the durable controls that hold.
When the Model Gets Recalled
On June 12, 2026 a federal export-control directive forced Anthropic to pull Fable 5 and Mythos 5 worldwide. My always-on agent was running on Fable 5. Here is how it died, why agentic systems have this hidden single point of failure, and what sovereign model-supply-chain risk means for anyone building on frontier models.
Experience
- Jan 2025 - Present
Founder & Agentic AI Solutions Consultant
Blackfrost Software CorpArchitect and operate an AI-native security ecosystem — offensive agentic-AI red-teaming and defensive multi-site SOC/EDR with AI-driven alert triage
- Aug 2022 - Jan 2025
IT Operations & Security Manager
Drive On Transportation LLCArchitected zero-trust network security across GCP, AWS, and Azure with Tailscale mesh networking
- Nov 2017 - Aug 2022
IT Infrastructure Lead & Systems Administrator
Simply Vegas Real EstateMaintained IT infrastructure for 50+ user organization with Salesforce CRM administration
- Apr 2014 - Aug 2015
Infrastructure Field Technician
Pulau Corporation (Fort Irwin / NTC)Maintained telemetry and networking infrastructure under DoD operational guidelines
- Jan 2004 - Jan 2012
Human Resource Operations Manager (NCOIC)
U.S. ArmyPreviously held a Secret clearance (eligible for reinstatement); managed PII for 1,500+ personnel with zero security breaches
Let's build something secure.
Federal, cleared, or private-sector engagements — book a consultation and we'll scope it together.
Book a consultation