FintechAI-Driven

DayTwo Mobile App

Built and shipped a production React Native fintech app helping SMEs track finances, manage debt, and get AI-assisted business guidance, with offline-first support across iOS and Android.

Senior Frontend Engineer5-person team1 yr 9 mosMobile
The Challenge

Most SME owners tracked business finances through notebooks or spreadsheets with no clear visibility into cash flow, expenses, or debts, making it difficult to make informed business decisions. The engineering challenge was building a production-ready fintech app with a 5-person team, covering AI-assisted guidance and offline reliability in low-connectivity environments.

By the time most SME owners knew they were losing money, it was already too late to act.

What I Did

Owned frontend architecture and mobile delivery for the DayTwo app, from system design through production release on iOS and Android.

Mobile architecture and state layer

Defined the React Native architecture using TypeScript, Zustand, and TanStack Query, establishing predictable data flow and reusable patterns that let the product scale without significant rework.

Core financial system

Built end-to-end financial workflows covering cash tracking, sales, expenses, debt management, and lending flows used by SME owners for daily business operations.

AI-assisted receipt capture

Shipped OCR-based receipt scanning that extracted transaction data from photos and mapped it into structured financial records, reducing manual entry for high-frequency users.

AI business advisor chat

Built a production LLM-powered chat experience with token streaming, structured response rendering, and retry handling, turning AI outputs into clear financial guidance for business owners.

Offline-first data sync

Designed local persistence with background sync to the REST API, ensuring users could record and review financial data reliably even without a stable internet connection.

Push notifications for lending workflows

Integrated Firebase push notifications for repayment reminders and lending status updates, keeping borrowers and lenders informed of time-sensitive lending activities.

Release pipeline and engineering standards

Built and owned iOS and Android release pipelines using EAS and GitHub Actions, enabling stable cross-platform delivery and consistent release workflows.

Results

800+

SME Users Onboarded

Businesses actively using the platform after launch

$37K+

Total Revenue Generated

Generated through lending and financial workflows

40%

API Calls Reduced

Reduced through TanStack Query caching and request deduplication

AI-powered

Business Advisor Chat

Production LLM chat with streaming responses and structured output rendering

Offline-first

Reliable on Any Network

Local persistence and background sync for low-connectivity environments

Multi-language

Localized Experience

Supported across multiple regions and languages

Visuals
What I Learned
  • 1

    Frontend architecture drives how fast the team can deliver. Early architectural decisions determine feature delivery speed and how sustainably the codebase scales as the product grows.

  • 2

    AI features require more than just connecting an API. Making the AI advisor feel trustworthy to users required handling slow responses, failed requests, and unstructured outputs before it was reliable enough to ship.

  • 3

    Offline support needs to be planned from the beginning. For SME users operating in low-connectivity environments, offline reliability is what makes the app usable in their daily operations.

  • 4

    Talking directly to users changes how you build. Working with SME users revealed workflow assumptions that changed how several core features were designed.

  • 5

    Automating releases early removes friction on every sprint. Setting up the build and release pipeline at the start meant the team could ship updates and fix issues without manual steps slowing things down.