Credit Card Platform Engineering

Repayments Service

The Challenge

The company needed to migrate away from their legacy repayments vendor to a modern, reliable payment infrastructure. This required integrating with Stripe while ensuring zero disruption to existing customers' scheduled payments.

The Solution

I built a serverless repayments service on AWS using Lambda functions and Hono for API routing. The system handles payment processing, webhook events from Stripe, and communicates with other services via EventBridge.

Key Technical Decisions

• AWS Lambda + Hono: Lightweight, serverless architecture for cost-effective scaling
• EventBridge: Event-driven communication between services for loose coupling
• Stripe Webhooks: Real-time payment status updates with idempotent event handling
• GitHub Actions: CI/CD pipeline for automated testing and deployment

Customer Onboarding

The Challenge

New customers needed a streamlined but compliant onboarding process that handles credit checks, regulatory requirements, and account setup in a single flow.

The Solution

I developed a multi-stage onboarding form that guides users through the entire process:

• Credit Assessment: Integration with credit check providers to determine eligibility
• Credit Decisioning: Automated calculation of available credit limit and interest rate
• Document Generation: PDF generation for terms and conditions
• Compliance: Digital acceptance of T&Cs with audit trail
• Security Setup: 2FA configuration during onboarding
• App Handoff: QR code generation prompting users to download the mobile app and log in with their configured PIN

Customer Support Chatbot

The Challenge

Call center costs were high, and many customer queries were repetitive questions that could be resolved without human intervention.

The Solution

I implemented an AI-powered customer chatbot using MCP (Model Context Protocol) integrated with the company's standard operating procedures. The chatbot can:

• Answer common questions about accounts, payments, and card features
• Guide customers through self-service actions
• Escalate complex issues to human agents with full context
• Follow compliance-approved responses from SOPs

This reduced call center volume while maintaining quality customer support.

Support Agent Analytics Dashboards

The Challenge

When customers did need human support, agents lacked a unified view of the customer's account, making it difficult to quickly understand and resolve issues.

The Solution

I built analytics dashboards that give customer support agents a comprehensive overview of the customer they're helping, including:

• Account status and history
• Recent transactions and payment activity
• Previous support interactions
• Relevant flags and alerts

This enabled agents to resolve issues faster and provide more informed support.

Automated Batch Communications

The Challenge

Customers needed timely notifications about important account changes, but these were being handled manually or through unreliable legacy systems.

The Solution

I automated batch jobs for customer communications, including:

• Credit Limit Changes: Notifications when credit limits are adjusted
• Interest Rate Changes: Alerts about upcoming rate changes
• PIN Resets: Automated PIN reset confirmations and instructions

These jobs run on schedule via AWS, ensuring customers receive consistent, timely communications about their accounts.

Results

Across these projects, I helped modernise the company's core platform infrastructure—replacing legacy vendors, reducing operational costs through automation, and improving both customer and agent experiences.