Engineering
Building This Site with Hermes Agent
From a blank Astro project to a deployed site in one session. Here's how an AI agent wrote the code, generated the content, and pushed to production.
Available for projects
Make it work.
Trust you can verify.
When the cloud drops, your systems shouldn't. When the audit comes, your data should be yours to account for. AI that works offline, leaks nothing, and proves it — built in, not bolted on.
Built Different
One builder. Two workstations. Local LLM inference on real hardware, not a cloud API with a quota. Everything is self-hosted, private, and runs offline.
The AI industry builds for connectivity. I build for what happens when it fails. Resilience isn't a feature you add — it's architecture you commit to. Your data doesn't leave the box. Your systems don't drop when the cloud does. Your trust is verifiable, not taken on faith.
🖥️
Primary Workstation
Intel i9-14900K · RTX 5070 Ti · 32GB DDR5 · Pop!_OS
🖥️
Secondary Workstation
Intel i7-12700KF · Dual RX 9060 XT · 32GB DDR4 · Linux Mint
🧠
Local AI Stack
Ollama · Qwen · Granite · Phi · Nemotron
What I Build
Resilient, private, verifiable — not features. Architecture.
🛡️
Resilience Engineering
AI systems that run when the cloud doesn't. Offline-capable agents, local inference, and fallback architectures that keep working under failure.
🔒
Privacy Architecture
Zero telemetry by design, not by configuration. Your data stays on your hardware. Audit every packet. Prove nothing leaves.
✅
Verifiable Trust
Trust measured in logs, not promises. Inspectable behavior, transparent operations, and proof that your systems do exactly what they claim.
🎯
Integration & Consulting
From audit to deployment. Find the gaps in your AI stack — cloud dependency, telemetry leaks, compliance risks — and close them by design.
Featured Builds
Projects shipping from the Workstation Duo.
AI Agent Platform Live
RixBot
Private AI chatbot with full R&D knowledge base access. Demo mode with zero cloud dependency, live mode via Ollama endpoint.
Astro · JS · RAG
AI Agent Platform Running
Hermes Agent Bot
Autonomous AI assistant deployed on Telegram. Manages coding tasks, research, and DevOps — 24/7.
Python · Telethon · Ollama · Docker
Personal Site Live
elect-rix.tech
This site — Astro, Bun, Tailwind. Built locally, deployed to Cloudflare Pages, generated with AI assistance.
Astro · Bun · Tailwind
Lab Infrastructure Active
Workstation Duo
Two-machine setup: local LLM inference, distributed builds, GPU compute, and self-hosted services.
Pop!_OS · Ollama · Docker
From the Blog
Experiments, tutorials, and lessons learned.