RBUFFY.LABROHAN / ENGINEERING SYSTEMS
← BACK TO THE LAB
LIVE EXPLORATION / PUBLIC DEMOS

Curiosity is
interactive.

A homelab is easier to understand when you can experiment with its ideas. Try the room-aware thermostat logic right now, then explore how the RTX-backed AI demo will work when its connection goes live.

AVAILABLE NOW · BROWSER-BASED

Room-aware thermostat explorer

Adjust three room temperatures and watch a weighted temperature controller choose a setpoint request. The interactive rules reflect my documented heating-controller project, including a bedroom heat limit and a deadband.

Try the thermostat simulator ↗

Works entirely in the browser. No account, hardware connection, or installation needed.

// THE ACTUAL PROJECT'S DESIGN
T = 0.2 × living + 0.4 × bedroom
+ 0.4 × roommate
Target: 69°F
Control band: 68.5–69.5°F
Bedroom limit: 72°F
// demonstration only; no device access
COMING NEXT · GPU CONNECTION

RTX AI inference playground

An interactive application backed by a model on the actual RTX 4060 Ti 16 GB workstation. The local interface and backend implementation are prepared; the visitor-facing connection will be enabled after deployment and access testing.

Public inference not online yet

In the meantime, read the demo architecture or the research platform case study.

// LOCAL MODEL DEMO ARCHITECTURE
Visitor → dedicated app
Application: Gradio / Python
Inference: Ollama on Buffy
GPU: NVIDIA RTX 4060 Ti · 16 GB
Limits: request size + concurrency
// external gateway not yet connected
RESEARCH

AI and MLOps

Explore how datasets, orchestration and artifacts fit around local compute.

See the workflow ↗
OBSERVABILITY

Understand the system

Learn how metrics, logs and alerts help run many services on a single host.

See the platform ↗
SYSTEMS

Explore the software

Filter the technologies running in the self-hosted ecosystem by role.

Browse the stack ↗