Overview
I designed and deployed a distributed private-cloud environment that connected multiple hardware nodes into a single secure mesh network using Tailscale on top of WireGuard. The goal was to build a flexible AI-oriented environment where different systems could specialize and still operate as one coordinated platform.
Architecture
The environment paired a high-performance Ubuntu AI server with an RTX 3070 for heavier inference work and a Raspberry Pi layer for orchestration and lightweight control tasks.
Instead of treating each machine as a standalone island, I designed the setup so the nodes complemented each other:
- A GPU-capable Ubuntu node for heavier AI tasks
- A Raspberry Pi orchestration node for coordination and lightweight automation
- Secure connectivity across the whole environment without exposing public-facing ports
Orchestration
I used n8n for workflow automation and OpenClaw to coordinate task distribution across the hardware nodes. That gave the environment a practical control layer for routing jobs, triggering actions, and managing cross-node workflows without requiring everything to live on one machine.
Security Model
Security was built into the architecture rather than layered on after the fact. I implemented a zero-trust approach so communication between nodes stayed encrypted and isolated from public internet exposure.
That meant:
- Encrypted inter-node communication
- Reduced public attack surface
- Cleaner segmentation between orchestration and inference roles
Why It Matters
This project reflects the kind of technical work I enjoy most: secure architecture, distributed systems thinking, networking, automation, and practical AI infrastructure. It is the clearest example on the site of how I approach systems design across multiple layers at once.