Fundthesis

AI-Powered Financial Education Platform
By: Ali Benrami and Hasnain Niazi

Overview

Fundthesis empowers first-time and seasoned retail investors with real-time news summarization and AI-powered stock analysis for portfolio diversification. Users gain access to tools offering live and historical, in-depth company insights.

Why It Matters

Fundthesis bridges the gap in financial literacy by simplifying portfolio selection. By unifying market data, sentiment, and model-driven insights, it helps users make informed decisions without juggling multiple information sources.

Features

• Real-time market news summarization
• AI-driven stock insights and portfolio diversification guidance
• Company deep-dives with historical and live metrics
• Theme toggle (light/dark) and modern UI with shadcn/ui

Tech Stack

• Frontend: TypeScript (React), Next.js (Better-Auth, shadcn/ui)
• Backend: Python (PyTorch, Pandas, NumPy, scikit-learn, XGBoost, Optuna, Joblib, Transformers, Sentence-Transformers, Google-GenerativeAI, LangChain, LlamaIndex, NLTK, TextBlob, TheFuzz, ChromaDB, Surprise)
• APIs: yFinance, Alpha Vantage, Twelve Data, Finnhub, IEX Cloud, NewsAPI, Marketaux, OpenFIGI, Financial Modeling Prep, SEC EDGAR API, Gemini, Cohere, Hugging Face Inference API, Plaid
• Database: PostgreSQL (for financial and user data)

AI Components

• Clustering (unsupervised): Segment users and stocks to power personalized guidance
• Time-series forecasting: ARIMA/Prophet with tree-based methods (e.g., XGBoost)
• NLP: Finance-tuned transformers (FinBERT, Sentence-BERT) + vector search for sentiment and summarization
• Deep learning: From MLPs to LSTMs/TabNet for mixed data types and robust risk profiles

Prerequisites

• Node.js (LTS recommended). If you use nvm:

source ~/.nvm/nvm.sh
nvm use --lts

Quick Start (Frontend)

Install dependencies and run the dev server (choose one package manager):

npm install && npm run dev
# or: yarn && yarn dev
# or: pnpm install && pnpm dev
# or: bun install && bun dev

Open http://localhost:3000 to view the app.

Project Structure

fundthesis/
  src/
    app/
      layout.tsx        # App providers (e.g., ThemeProvider) and base layout
      page.tsx          # Landing page
      globals.css       # Global styles (Tailwind v4)
    components/
      ThemeButton.tsx   # UI for theme switching
      ui/               # shadcn/ui components (e.g., button)
  public/               # Static assets
  README.md

Environment Variables

Create a .env.local file in the project root for local development. Examples:

NEXT_PUBLIC_API_BASE_URL=http://localhost:8000
BETTER_AUTH_SECRET=replace_me

Development Notes

• Styling uses Tailwind CSS v4 and PostCSS. See src/app/globals.css and postcss.config.mjs.
• App layout and providers live in src/app/layout.tsx.
• Main page: src/app/page.tsx.
• UI components under src/components/.

Roadmap (Excerpt)

• Weeks 2–3: Data collection, cleaning, and storage (stocks + budgeting)
• Weeks 4–5: Unsupervised clustering (users and stocks) + visualization
• Weeks 6–7: Predictive modeling (stock scoring, budget forecasting)
• Weeks 8–9: Frontend integration + RAG assistant (LangChain + vector DB)
• Weeks 10–11: Final polish, presentation, and demo

See the full plan in Fundthesis White Page.md.

Troubleshooting

• If themes don’t switch: ensure next-themes is installed and ThemeProvider wraps the app in src/app/layout.tsx with attribute="class".
• If Node/npm not found: install Node LTS or use nvm (nvm use --lts).

Contributing

1. Clone the repo and create a feature branch
2. Make changes with clear commits and open a PR
3. Ensure linting passes and add concise documentation where relevant

License

TBD