therml

Thermal ∩ ML — workloads migrate so hardware can rest.

The therml Philosophy

The Approach

Instead of forcing HVAC systems to work harder, `therml` uses a real-time AI engine to read the thermal map of the data center grid.

It dynamically migrates virtual workloads to physically cooler nodes—letting hot hardware dissipate heat naturally.

The Impact

Built for the Google Solution Challenge to drastically reduce Scope 2 emissions in digital infrastructure.

SDG 9: Infrastructure SDG 13: Climate SDG 7: Energy

Team Icarus

⚙️
Shashank Kontikal
Team Lead
"Orchestrating the perfect sync."
🧠
V Muhammed Maazin
ML Engineer
"Teaching AI to feel the heat."
📊
Sharvesh Raj
UX Vis
"Making data look cool."
⚡
Umar Farroq
Backend Optimize
"Squeezing out latency."

Test the Shift

// Instructions

1. Open the Simulation Dashboard to create racks and inject heat.

2. Open the Monitoring Dashboard to view the live thermal heatmap.

3. Watch the AI autonomously migrate workloads from hot to cool.

4. Try a manual Override to pin work to a hot rack and see the AI adapt.

Open Monitor Open Simulator

Get the Source

`therml` is open-source. Explore the architecture.

View on GitHub