WEHack2026

Team

Areej Fatima

Hanna Kellenberger

Overview

We built MemoryOps, an AI system that helps real estate teams learn from past building events to make better decisions. Instead of forgetting data like most systems, it stores incidents like maintenance issues and tenant complaints as “memories,” finds patterns over time, and uses them to explain risk and recommend actions. It works like a living museum of building history—helping users understand what went wrong in the past so they can prevent problems in the future.

Tech Stack

Frontend: HTML, CSS, Javascript
Backend: Python(flask)
Database: SQLite
Visualization: Charts.js
AI integration: Google Gemini API

Features

• Incident logging - Log building events(water leaks, AC failure, HVAC faults, cybersecurity alerts, outages) and log by severity.
• Risk Analysis - Scores buildings based off of repeated issues, severity trends, and category based risks.
• AI-Power Insights - Ask Gemini questions about your buildings. "Which building is most at risk?" or "What patterns are emerging?” Gemini will respond back only with infomration it has retrieved from the database to get rid of hallucinations.
• Memory System - Builds structured memory per building and stores historical incidents.
• Portfolio Intelligence - Generates portfolio reports that include and executive summary, risk rating, and saves previous analyses for comparison.