A quantitative model to extract signals from noise
Context
This project was the part of a Hackathon conducted by Artificial Intelligence Society at University of Texas at Dallas.
Role
UX Designer
Team
1 designer (me),
4 AI engineer
Duration
24 hours
Awards
First in Figma Track
Second in Quants AI Track
Finding a needle in a haystack
Problem
Analysts and investors spend hours connecting news, executive statements, and filings, yet key insights still go unnoticed.
Solution
Our system links this text-heavy data to stock price movements to detect contradictions. It flags positive, negative, or watch signals and tracks signal decay. We also showed users the statistical evidence behind each signal.
