Inspiration

Most financial tools show data but don’t help you decide what to do next. I realized the real gap isn’t data — it’s decision-making. So I built CapitalAI to answer a simple question: What if your dashboard could think, predict, and guide you like a financial advisor?

What it does -> CapitalAI is an AI-powered decision platform that helps businesses: ~Predict future ROI ~Analyze mergers and suggest decisions ~Compare performance with real market data ~Optimize capital allocation using simulations

Instead of just showing numbers, it tells you what will happen, why, and what to do next.

How we built it

Backend: FastAPI with modular AI agents ML Models: XGBoost for ROI prediction using time-series data Explainability: SHAP for understanding model decisions Database: MySQL with real financial + stock data Frontend: React with a clean, Plaid-inspired UI Deployment: Railway

Challenges we ran into

~Turning ML predictions into clear decisions ~Making AI outputs explainable and trustworthy ~Keeping data realistic across multiple companies ~Balancing complex logic with a simple UI

Accomplishments that we're proud of

~Built a full end-to-end AI product, not just a model ~Created a working merger prediction system ~Combined ML, explainability, and simulation in one platform ~Delivered a clean UI that shows real decision impact

What we learned

~How to build AI systems, not just models ~Why explainability is critical in real-world AI ~How to connect ML with business decision-making ~Full-stack integration of AI into a usable product

What's next for CapitalAI

~Real-time financial data integration ~Stronger RL-based optimization ~More advanced simulations and what-if scenarios ~Smarter AI explanations using LLMs