Overview

Judging is based on four criteria, 25% each:

1. Relevant Sustainability Challenge Addressed

2. Creativity

3. Technical Feasability

4. Pathways for Scale

Each criterion is scored 0–10 points.
Total possible score: 40 points (each criterion weighted equally).

Description:
How clearly and effectively does the project tackle a pressing environmental or sustainability issue?

Guiding Questions:

• Does it fit the Green-in AI (making AI more efficient) or Green-by AI (using AI for sustainability) challenge?

• Does it demonstrate a measurable sustainability benefit (e.g., reduced compute intensity, carbon savings, improved climate decision-making) demonstrated with metrics and statistics?

• Does it identify a specific problem and propose a realistic technical pathway toward addressing it?
  
  

Weight: 25%

Description:
We want hackers to create a project that makes you say “wow.”

Guiding Questions:

• Is the idea novel, surprising, or presented in an unexpected way?

• Does it think beyond conventional boundaries, reimagining how AI or sustainability problems are approached?

• Is there evidence of original thought in concept, design, or user experience?
  
  

Weight: 25%

Description:
In only 22 hours, hackers manage to build projects with remarkably complex infrastructures. Despite this crunch, we hope to see beautiful code or hardware under the hood.

Guiding Questions:

• How technically challenging was the implementation?

• Did the team build or integrate advanced systems (e.g., ML models, geospatial pipelines, API orchestration, hardware)?

• Is the code or architecture efficient, scalable, or elegant?

• Does the demo actually work and showcase functional technology?
  
  

Weight: 25%

Description:
How feasible and scalable is the solution beyond the hackathon?

Guiding Questions:

• Could the solution be deployed, extended, or adopted by real users, organizations, or policymakers?

• Does it outline next steps, partnerships, or measurable impact metrics?

• Would further development amplify sustainability outcomes?
  
  

Weight: 25%

Green-in AI (Powered by Microsoft)
Projects that:

• Reduce compute or energy use per task (e.g., carbon-aware scheduling, adaptive scaling).

• Increase the sustainability value delivered per unit of compute (e.g., model compression, efficient inference).
  
  

Green-by AI (Powered by Google)
Projects that:

• Use Earth Engine or other geospatial data to monitor, predict, or optimize for sustainability.

• Apply AI to climate, agriculture, energy, or resource-management use cases (e.g., detecting deforestation, forecasting droughts, optimizing logistics).