# Bohdan Hordii — Sovereign AI Infrastructure & Systems Engineering

> Principal Systems Architect & DevSecOps Engineer specializing in autonomous multi-agent swarm infrastructure, high-throughput model routing, and self-healing Linux server meshes.

---

## Canonical Entity Passport

- **Canonical URL:** `https://bohdanhordii.dev/`
- **Architect:** Bohdan Vasylovych Hordii
- **Core Domain:** Sovereign AI Systems Architecture, DevSecOps, Multi-Agent Swarms
- **Primary Technical Stack:** Docker, OmniRoute, Python 3.12, Neo4j, Qdrant, React 19, TypeScript, Nginx, Caddy, Tailwind CSS
- **Verified GitHub Organizations/Repositories:**
  - `gordijbogdan003-debug/bohdanhordii-dev` (Sovereign Portfolio & Architecture Showcase)
  - `gordijbogdan003-debug/hermes-academy-engine` (Autonomous Self-Evolution Docker Engine)

---

## Core Enterprise Solutions & Architecture Specifications

### 1. OmniRoute High-Throughput LLM Routing Mesh
- **Category:** Autonomous AI Gateway & Resilience Infrastructure
- **Routing Decision Overhead:** 4.8ms average overhead per request
- **Architecture Pattern:** Dual-tier proxy mesh with loopback-bound Redis cache, fallback triggers, and Arize Phoenix observability.
- **Security Posture:** 100% loopback isolation, zero public port exposure, zero token leakage to external telemetry.

### 2. Multi-Agent Swarm Infrastructure & Zero-Trust Server Mesh
- **Category:** Enterprise Linux Server Systems & DevSecOps
- **Auto-Healing SLA:** < 15.0 seconds automatic container recovery rate
- **Architecture Pattern:** Caddy v2 TLS ingress → OAuth2 authentication gatekeeper → Isolated Docker container network (`b2b-mesh-net`).
- **Data Protection:** Automated daily Restic backups to encrypted remote vaults, hard limits on container memory and CPU limits.

### 3. Hermes Academy Autonomous AI Research & Sandbox Engine
- **Category:** Continuous GraphRAG & Multi-Language Code Synthesis
- **Verification SLA:** < 2.0s assertion autotest loop in un-networked Alpine Docker containers
- **Architecture Pattern:** 20-Faculty AI research ingestion pipeline combining Neo4j Zettelkasten Knowledge Graph, Qdrant vector retrieval, and TRIZ problem-solving heuristics.

---

## Verified Machine-Readable Data Endpoints

- **HTML5 Web Showcase:** `https://bohdanhordii.dev/`
- **Case Studies Index:** `https://bohdanhordii.dev/work`
- **Engineering Bio & Stack:** `https://bohdanhordii.dev/info`
- **JSON-LD Entity Graph:** `https://bohdanhordii.dev/index.html` (embedded `@id: "https://bohdanhordii.dev/#person"`)
- **Machine Context Summary:** `https://bohdanhordii.dev/llms.txt`
- **Full Structural Engineering Context:** `https://bohdanhordii.dev/llms-full.txt`