Skip to main content ↓
Back to Projects
fintech • 2022 — Present
SGeBIZ (Singapore & Malaysia) • Senior Software Engineer (Frontend Lead) • Technical Consultant

Ezypayment 2.0 Fintech & Security Architecture

Rewriting a corporate credit card payment platform to pass all enterprise bank security audits

Ezypayment is a corporate payment platform by SGeBIZ that lets businesses make vendor payments using commercial credit cards. As Frontend Lead and continuing Technical Consultant, I led the complete rebuild using clean architecture, dropping crash rates from 3.8% to under 0.1% and passing all third-party banking security penetration audits.

<0.1%
Production Crash Rate
100%
Bank Audits Passed
-35%
Bundle Size Reduced
50%
Faster QA Pipeline

Ezypayment 2.0 Fintech & Security Architecture — System Topology

Bank-grade layered financial architecture enforcing Clean Architecture, encrypted isolate communication, SSL certificate pinning, and hardened payment tokenization.

Tap node to inspect
client

Ezypayment UI Layer

(Flutter Reactive Widgets)
Zero Coupling

Declarative UI layer strictly observing domain state with zero direct network calls.

Data Flow: Client Layer → Gateway → Domain → Cloud
Client Apps Edge Gateway Microservices Third-Party / Banking
4 Active Protocol Contracts
Interactive Architecture Simulation

Bank-Grade Corporate Invoice Payment Simulation

Trace secure corporate credit card processing through enterprise bank audit layers

Security 0ms

1. Card Entry & Enclave

Hardware keystore encryption & root/jailbreak detection

Pending
Mobile +50ms

2. Clean Arch Domain

Pure Dart business layer with zero framework coupling

Pending
Network +65ms

3. SSL Pinning Bridge

Strict certificate pinning against MITM intrusion

Pending
Bank API +120ms

4. Bank Core API

Corporate credit authorization & 2FA transaction hash

Pending
Audited +50ms

5. Audit Verified

100% penetration test pass & encrypted audit log

Pending
Target Latency: 285ms • Reliability: <0.1% Crash Rate SLA
Click "Run Architecture Simulation" to test pipeline

Challenges & Engineering Decisions

How key technical problems were solved in production

Challenge 1
Problem

The legacy app mixed UI code with business logic, causing state bugs and frequent edge-case crashes.

Design Decision

Separated the app into distinct Presentation, Domain, and Data layers with 100% pure Dart business logic.

Production Impact • Dropped crash rates below 0.1% and achieved 90%+ unit test coverage on payment calculation logic.
Challenge 2
Problem

Banking partners required strict security: no plain-text memory leaks, jailbreak and root detection, and encrypted storage.

Design Decision

Implemented hardware-backed secure storage (Keychain and EncryptedSharedPreferences), dynamic SSL pinning, and root detection.

Production Impact • Passed all third-party banking security assessments with zero high-severity findings.
Challenge 3
Problem

Managing Dev, UAT, Staging, and Production builds manually was slow and prone to configuration mistakes.

Design Decision

Built automated multi-flavor CI/CD build scripts with dynamic bundle IDs and signing certificates.

Production Impact • Cut release preparation time by 50% and eliminated configuration errors.

Quantifiable Performance Benchmarks

Empirical metric delta measured before and after architectural refactor

Crash-Free User Rate
Baseline: 96.2%
Result: 99.92%
+3.72% Stability
App Binary Size (iOS)
Baseline: 78 MB
Result: 49 MB
-37% Smaller
CI/CD QA Build Time
Baseline: 45 mins manual
Result: 18 mins automated
60% Faster
Penetration Testing Defects
Baseline: 14 initial flags
Result: 0 (Clean report)
100% Remediation

PRODUCTION TECHNOLOGY STACK

8 Technologies
Flutter
Dart
Clean Architecture
Bloc State Management
SSL Pinning
Secure Enclave / Keychain
Fastlane CI/CD
Banking Gateway APIs