JAMI_SpendWise

Budgeting UI. Which helps to categorised live data by pre-trained model
Goal planner
ML based goal tracker with predictions and recommendations
Main dashboard with aggregation of monthly data and projections

JAMI is a smart AI-powered assistant designed to help people achieve their financial goals - one step at a time. It was born from a simple story: a dream trip to Japan that never happened, year after year, because of unclear budgets and unnoticed overspending. JAMI steps in as both guide and partner, helping users plan realistically and stay committed.

You can create personal goals - like travel or retirement - and JAMI will help break them into manageable monthly targets. As your financial activity evolves, JAMI offers real-time alerts if you’re about to overspend and suggests ways to stay on track. It connects emotional motivation with smart financial planning, blending story and strategy. More over it help to extrapolate and predict your future accumulated balance.

Challenges we ran into

We challenged ourselves to go beyond a basic demo - our goal was to build a working prototype that reflects the full product vision.

A key challenge was integrating two complementary AI approaches into a seamless experience:

Prediction Agents: Focused on forecasting financial trends and projecting future scenarios.

Budget Optimisation Agents: Designed to analyse spending habits and suggest efficient improvements.

Merging these systems required both architectural planning and practical experimentation. On top of that, we stepped far outside our technical comfort zones. While our backgrounds are mainly in backend development, we built the entire product in JavaScript, including interactive UI features.

We also took on the full AWS ecosystem from scratch - setting up cloud infrastructure, hosting, storage, and services - all without relying on pre-made templates. It was a steep learning curve, but an incredibly rewarding one.