Judy | Devpost

Inspiration

The inspiration for Judy stems from a user standpoint of view; people want to save money, but they don't want to agonize over their utility bill or constantly pay attention to what time they're using their regular devices. By recognizing user patterns and automating affordable-charging times, Judy increases economic incentives and removes external effort.

What it does

Judy connects smart plugs to servers in a digital app to keep track of devices' electrical usage, especially during peak hours. Its AI capabilities learn from users' usage habits and recommends pausing electric usage for certain devices during peak loads, saving customers money and helping utilities manage electric spikes.

How we built it

We built using Python and Mako, along with AI reinforcement learning. We also combined it with the hardware of the smart plug for our demonstration.

Challenges we ran into

One main challenge in the beginning was connecting the hardware with the computer server, while also pairing both with a phone to begin building the foundation of Judy.

Accomplishments that we're proud of

We are proud of learning how to create the AI model behind Judy, managing the hardware of the smart plug, and figuring out how to show its capabilities in the demo.

What we learned

We learned how to connect the hardware of the smart plug to the computer server, as well as how to build an AI model that can not only keep track of devices, but know when to pause charging to reduce prices for users.

What's next for Judy

What might be next for Judy is to create a pilot program for a small, localized area to see how a network of smart plugs might be managed. Furthermore, developing a voice agent for Judy could minimize users' need for constantly interacting with a website or app.