Problem Statement: Traditional investment advising often lacks sector-aware, diverse perspectives and fails to dynamically integrate real-time public sentiment. Investors are overwhelmed by mixed signals and struggle to make informed decisions. The challenge was to create a solution that provides clear, actionable guidance while reducing cognitive overload for retail investors and advisors.
Solution: Developed Quorum, a sector-aware solution to investment advising. It uses a council of LLM advisors to deliberate on a stock ticker with its provided data from Yfinance. Additionally, it features a customer sentiment advisor that uses the Alpha Vantage API to determine the public sentiment on the asset.
Skills Used:
Conceptualized the platform in collaboration with Persistent Innovation Labs and MSOE AI Club. The primary goal was to build a tool that turns noisy financial signals into decision-ready guidance, reducing cognitive overload for retail investors and advisors.
Developed the Investment Council dashboard, allowing users to compare multiple stock tickers side-by-side (such as MSFT and NVDA). The dashboard generates a council confidence score and clear Buy, Hold, or Avoid ratings powered by a specialized team of AI advisors (Risk, Growth, Fundamentals, Sector Strategist, and Sentiment).
Ensured the system was 'Auditable by Design' to avoid black-box AI outputs. Each advisor provides transparent reasoning modals displaying underlying ML model signals, conviction percentages, and key financial metrics like log-returns, volatility, and moving average outperformance.
Architected the software to be 'Fit for Industry', structuring data ingestion pipelines to easily support plugins for industry-standard platforms like Bloomberg, PitchBook, and S&P Global.