marketmind - CalHacks 11.0
MarketMind
At CalHacks 11.0, my team and I created MarketMind, an AI-powered platform that simulates public responses to media campaigns before release. We built a multi-agent system to models human interaction and opinion evolution to help companies predict and refine their content. Our platform enables businesses to test multimedia content against diverse user personas, providing real-time analytics and visualization of sentiment progression before releasing possibly controversial marketing campaigns.
My Role
As a full-stack engineer, I led the development of our real-time analytics dashboard using React and TypeScript after creating a high-fidelity mockup in Figma. I then integrated Google Gemini and Fetch.ai to create realistic user personas and facilitated agent interactions through the SingleStore database for data processing.
Key Features
Multimodal Content Testing
Test any form of media content (video, text, audio, images) against AI-generated personas that represent your target demographic using Google Gemini and Fetch.ai's agent system.
Dynamic Sentiment Analysis
Monitor how opinions evolve over time through multiple simulation cycles, with real-time visualization of sentiment shifts and potential controversy points.
Intelligent Agent Interactions
Leverage sophisticated agent-to-agent interactions to model realistic human behavior and conversation patterns within target markets.
Real-Time Analytics Dashboard
View comprehensive analytics through dynamic visualizations built with React and Recharts, offering instant insights into campaign performance and potential risks.
The platform was built using Google Gemini for persona generation, Fetch.ai for multi-agent simulation, SingleStore for real-time data processing, and React/TypeScript for the frontend interface.
Frames from the deck:
