Inspiration
Credit cards are powerful financial tools, yet most people don’t actually optimize them. Users often choose cards based on branding or habit, not reward efficiency, risk exposure, or long-term financial impact.
We asked a simple question:
What if every purchase triggered an intelligent financial decision engine?
FitFin was built to transform everyday transactions into optimized financial strategies.
The Problem
Commonly, people:
- Don’t know which card is best for a purchase
- Don’t understand how purchases affect long-term goals
- Rarely simulate financial risk (overdraft probability, balance drift)
- Leave reward value on the table
- Don’t know when it’s worth applying for a better card
Most financial tools focus on budgeting — not strategic credit optimization.
What FitFin Does
FitFin is an AI-powered financial decision engine that:
- Analyzes receipts using OCR
- Detects purchase categories
- Optimizes credit card selection for maximum net rewards
- Simulates 30-day balance outcomes using Monte Carlo modeling
- Evaluates financial risk impact
- Suggests smarter wallet upgrades when better reward opportunities exist
It doesn’t just tell you which card to use.
It tells you whether your wallet strategy itself needs upgrading.
How It Works
Receipt Intelligence
We extract structured purchase data (merchant, category, total) and normalize it into standardized reward categories.
Reward Optimization Engine
For each purchase:
- We compute reward rates across selected cards
- We calculate expected reward gain
- We apply a risk penalty adjustment
- We select the card that maximizes net score
[ score = rewards - risk_penalty ]
Monte Carlo Balance Forecast
We simulate 30-day balance trajectories using probabilistic spending assumptions:
- Baseline P10 / Median / P90 projections
- Overdraft probability estimation
- What-if emergency expense modeling
This allows users to see how a purchase shifts financial stability — not just rewards.
AI Upgrade Opportunity
If a non-selected available card offers a higher reward rate for the detected category, we compute:
[ \Delta = (bestAvailableRate - bestSelectedRate) \times purchaseAmount ]
If positive, FitFin recommends applying for that card.
This transforms the system from a selector into a financial strategist.
Challenges We Faced
- Designing an optimization layer that felt intelligent while remaining deterministic
- Integrating Monte Carlo simulation into a fast client-side experience
- Maintaining robustness while supporting future cloud-based optimizers (AWS extensibility)
- Making financial modeling understandable to non-technical users
Why FitFin Is Different
Most financial apps:
- Track spending
- Categorize expenses
- Provide dashboards
FitFin:
- Optimizes decisions in real time
- Simulates future financial risk
- Quantifies opportunity cost
- Suggests strategic credit upgrades
It combines reward optimization, probabilistic risk modeling, and scenario simulation into a single intelligent financial decision system.
In short:
FitFin doesn’t just track your finances — it actively optimizes them.
Built With
- amazon-web-services
- client-side-heuristic-scoring-system
- cloud
- custom-monte-carlo-simulation-engine
- deterministic-reward-optimization-algorithm
- fastapi
- javascript
- next.js-14
- numpy
- ocr-receipt-parsing
- probabilistic-financial-risk-modeling
- pydantic
- pyomo
- python
- react
- tailwind-css
- typescript

Log in or sign up for Devpost to join the conversation.