20+ years architecting secure, enterprise-grade systems — now building agentic AI for homes & businesses.
SurgeXi Business Intelligence (surgexi.com) — secure, human-in-the-loop AI operations software. Our lead product, Vertirite, is a vendor-neutral AI-governance control plane: it discovers every AI and automation already touching your systems and puts a human approval gate — plus egress, credential, and irreversible-action chokepoints — in front of the actions that can hurt you. Everything is built on a security-first, on-prem foundation: prompt-injection hardening, agent isolation, private GPU inference, and end-to-end CI/CD security review.
- Agentic AI orchestration — multi-tier LLM inference routing, fallback, and cost optimization
- Security architecture — prompt-injection defense, agent isolation, IAM, threat detection
- On-prem & edge AI — private GPU inference, low-latency model serving, airborne edge compute
- Industrial systems — MES / OT-IT integration on live automotive production floors
Each repo is a case study — architecture diagrams, design-decision records (ADRs), and a runnable reference implementation. They distill patterns from systems I've built in production; no proprietary code is exposed.
| Repo | What it shows |
|---|---|
| 🔎 verified-rag | RAG that structurally cannot fabricate a citation — every quote is verbatim-verified against source or the engine abstains — architecture & case study |
| 🧭 llm-inference-router | Health-aware, cost-optimized tiered LLM routing — architecture & case study |
| 🖥️ onprem-ai-infra | Private on-prem dual-GPU inference platform, infra-as-code — architecture & case study |
| 🛡️ secure-agent-patterns | Defensive patterns for hardening agentic AI — architecture & case study |
© 2026 SurgeXi Business Intelligence, a Teamsmith Enterprises LLC company. All Rights Reserved. Source-available for evaluation only — see LICENSE.