Competition Details
Competitions Screen
Home Screen
Gemini AI Portfolio Analysis

QuantQuest

Inspiration

Every college student hears the same advice: "Start investing early."

But when you open a brokerage app for the first time, you're faced with walls of numbers, unfamiliar terms like Sharpe ratio and mean reversion, and the fear of losing real money.

We asked ourselves: what if learning to invest felt more like playing a game than reading a textbook?

We were also inspired by the Capital One prompt: "How will you reimagine banking?" We realized the biggest barrier for young investors is the gap between having money in a bank account and knowing how to invest it.

QuantQuest bridges that gap by turning a simulated bank balance into a sandbox for learning investing with real market data, AI coaching, and blockchain-verified trades.

What it does

QuantQuest is a gamified investing platform where users start with a \$10,000 simulated bank account (powered by the Capital One Nessie API) and build a portfolio across three asset classes.

Asset Classes

Sectors

Weighted baskets of real stocks with live price feeds from Yahoo Finance.

Examples include:

Technology\
Finance\
Energy\
Healthcare\
FAANG+

Quantitative Strategies

Algorithmic investment strategies such as:

Momentum\
Mean Reversion\
Statistical Arbitrage\
Risk Parity

Each strategy has a realistic risk/return profile derived from historical behavior.

Safe Assets

Low-risk instruments designed to teach capital preservation:

Treasury Bonds\
High-Yield Savings\
TIPS\
CD Ladders

Gamified Learning

Users earn XP and achievements for good investing habits:

Diversifying portfolios\
Running AI portfolio analysis\
Reaching portfolio milestones

The progression system includes five levels, from Paper Trader → Quant Master.

AI Portfolio Analysis

Users can run AI-powered portfolio analysis via Google Gemini, which returns three perspectives:

Critic --- highlights risks and weaknesses\
Optimist --- identifies growth opportunities\
Realist --- balanced evaluation

Blockchain Transparency

Every trade is recorded on the Solana devnet, creating an immutable ledger of simulated trades.

Advanced Analytics

QuantQuest computes institutional-grade metrics such as:

Sharpe Ratio\
Sortino Ratio\
Maximum Drawdown\
Value-at-Risk (VaR)

How we built it

Frontend

Next.js 16\
React 19\
Tailwind CSS\
Recharts for portfolio visualizations

Backend

Next.js API routes\
MongoDB Atlas with Mongoose\
Dual persistence architecture (MongoDB primary + in-memory fallback)

This design allows the system to degrade gracefully if services fail.

Financial Modeling

Daily returns are simulated using volatility scaled by asset class:

$$ r_{\text{daily}} = \frac{r_{\text{annual}}}{252} + \epsilon, \quad \epsilon \sim U(-\sigma, \sigma) $$

Portfolio analytics use standard quantitative finance formulas.

Sharpe Ratio:

$$ \text{Sharpe} = \frac{\bar{r} - r_f}{\sigma}\sqrt{252} $$

Sortino Ratio:

$$ \text{Sortino} = \frac{\bar{r} - r_f}{\sigma_{\text{down}}}\sqrt{252} $$

Value-at-Risk:

$$ \text{VaR}_{95} = \bar{r} - 1.645\sigma $$

Historical simulations replicate events like the 2008 financial crisis and the COVID-19 market crash using trajectory-based price paths.

Challenges we ran into

Intraday portfolio accuracy

Refreshing portfolios multiple times per day risked compounding market changes incorrectly. We solved this using a daily baseline snapshot so changes are applied relative to the market open.

Blockchain latency

Solana confirmations take several seconds. To keep the UI responsive we implemented asynchronous trade recording.

External API reliability

The system depends on multiple external APIs (MongoDB, Solana, Gemini, Yahoo Finance). We implemented fallback paths for every integration so the app continues functioning if services fail.

Realistic market simulations

Market crashes do not move linearly. We implemented trajectory models that mimic real patterns like rapid crashes followed by gradual recovery.

Accomplishments that we're proud of

Meaningful integration of five sponsor APIs
Institutional-grade portfolio analytics accessible to beginner investors
Six playable historical market scenarios
A full gamification system with levels and achievements
A resilient architecture that continues working even if multiple services fail

What we learned

Quantitative finance is hard to simulate realistically. Modeling volatility, trading days, and compounding required careful design.
Blockchain systems require asynchronous architecture. Treating distributed systems as unreliable improved system resilience.
AI prompting is an engineering discipline. Producing structured multi-perspective analysis required careful prompt design.