Finpath

Problem

People rarely know where they stand financially, and generic advice ignores their actual situation. The product had to take what a person entered about their money and turn it into a real profile, then recommend concrete actions for that profile rather than the same tips for everyone.

What I built

• A profiling flow that collects the inputs that matter (income, spending, debt, goals, risk) and turns them into a financial profile, weighting the variables rather than treating them as a checklist.
• A calculation layer in Google Sheets, connected to Bubble, so the financial formulas live where they can be read, audited and adjusted without touching the app. Finance logic changes often; keeping it out of the interface means a change is a spreadsheet edit, not a redeploy.
• Recommendations driven by the profile: the actions a person sees follow from their own numbers and update when those numbers change, so the advice stays honest as their situation moves.
• A data model built for the questions behind the advice, so the app can reason about a person's finances over time rather than only showing today's snapshot.

Results

The interesting part of this one was the split: Bubble for the experience, a separate calculation layer for the maths. It kept a domain that changes constantly (financial rules and thresholds) out of the part that is expensive to change, which is exactly the data-and-logic-first approach I bring to every project.