MVP for a Retail Investment Advisor Powered by LLMs
Contributers: Yves D'hondt, George Sotiropoulos, Rishi Kumra, Naman Singhal
- Key Problems Tackled
- Key Tools Used
- Product Demo
- Pitch Deck
- MVP Screenshots
- Large amount of textual data
- News data
- Dense financial reports
- Large amount of numerical data
- Financial time series are notoriously noisy
- Requires good understanding of statistics & time series methods
- Quant finance models are gate kept
- The average retail investor does not have the resources or know-how to utilize advanced recommendation & asset allocation models
- Python
- GPT (current LLM used in the product)
- langhcain (support with large context windows)
- streamlit (user interface)
- pandas/numpy/plotly/seaborn (data visualization)
- (TBD) scikit-learn/pytorch (needed in the future to suport dynamic quant models)