C1G9_Sorite gang

Team Member Name

Chen Yong Jeen (Leader) Teoh Yu Heng Teoh Jun Ze Ng Xian Yang Saw Yong Quan

Inspiration

Malaysia’s rising electricity costs and increasing grid stress highlighted a real problem: households lack visibility and control over their energy usage. Many people only realize their consumption habits when the bill arrives—too late to change anything. We wanted to empower users with real-time insights, smarter decisions, and AI-powered guidance to reduce waste and save money.

What it does

PowerPulse is an AI-powered energy management platform that helps households monitor, analyze, and optimize electricity usage. Users can track their budget, understand appliance-level consumption, detect phantom loads, and receive personalized recommendations on when and how to use appliances efficiently based on demand, peak/off-peak pricing, and weather conditions.

How we built it

Frontend: Next.js 16, TypeScript, and Tailwind CSS for a fast, responsive UI

Backend & Database: Supabase (PostgreSQL + Auth) for real-time data storage and secure access

AI Planner: Groq SDK with Llama 3.3 70B to generate usage optimization plans

Phantom Detection: Python Flask API with ML model analysis

Visualization: Recharts for dynamic charts and usage trends

UI Components: Shadcn UI + Lucide icons for a polished design

Challenges we ran into

Limited access to real smart-meter data

Balancing real-time data performance with visual analytics

Designing AI recommendations that feel actionable and practical, not generic

Integrating multiple services (Next.js + Supabase + Flask + Groq) reliably

Ensuring accurate appliance cost segmentation without NILM hardware

Accomplishments that we're proud of

Working AI planner that adjusts usage based on user preferences

Early prototype phantom load detection model

Clean and responsive UI with smooth user experience

A solution that feels practical for real households, not just a concept

What we learned

Integrating real-world energy datasets is harder than expected

The importance of user-centric design when working with technical insights

How to build scalable cloud services and connect multiple API layers

Practical ML model deployment using Flask and Groq LLM integration

What's next for C1G9_Sorite gang

Deploy Production Systemby hosting on cloud platforms (Vercel + Render)

Enhance AI Chatbot/ Integrate Claude API for conversational recommendations to add voice interface capabilities and provide seasonal usage optimization tips

Expand Appliance Library to add more appliances (heater, kettle, rice cooker, iron, hair dryer)

Improve multi-state appliance detection (dishwashers, washing machines)

Real-Time Integration

Business Model Development

Regional Expansion

Partner with TNB or smart meter providers