Case Study: MIT Sustainable Design Lab accelerates urban energy modeling with Lightning AI

MIT Sustainable Design Lab - Customer Case Study

The MIT Sustainable Design Lab faced significant challenges in developing an application to model urban building emissions. Their team struggled with extremely long training loops that took over 30 hours per epoch and were slowed down by cumbersome boilerplate code and difficult environment setup, which blocked junior members for weeks and reduced overall productivity. They turned to Lightning AI and its product, Lightning Studios, to address these issues.

By implementing Lightning AI's solutions, the lab achieved substantial results. Using Lightning Trainer eliminated boilerplate code and automated logging integrations, while Lightning Studios provided powerful multi-GPU machines. This reduced model training time by 85% from 36 to 6 hours, cut prototyping time by 60%, and slashed new researcher onboarding time by 90%, dramatically accelerating their path from prototype to production.

MIT Sustainable Design Lab

Sam

Researcher