Back to Projects

Finance Case-Manager

A case platform that went from MVP to 18M+ monthly cases and 220M+ monthly transactions across multiple tenants.

Problem

We needed a new case-management foundation that could ship fast, handle a lot of traffic, and support queries DynamoDB was starting to fight us on.

Approach

I built the Spring Boot base, the stream processing pipeline, the encryption layer, and the deployment flow across PostgreSQL, DynamoDB, Kinesis, DMS, S3, and EKS.

Impact

We shipped the MVP in 8 weeks, pushed Kinesis to 150 TPS, cut encryption latency by 60%, and took sandbox deploy and test cycles from 20 minutes to 2.

Tech Stack

Spring Boot Next.js PostgreSQL DynamoDB Kinesis AWS KMS Terraform EKS Docker

Overview

Finance Case-Manager was the backbone for a new financial operations workflow. It had to move fast like an MVP, but it also had to survive production traffic and changing requirements.

System constraints

  • Multi-tenant traffic with large swings in transaction volume.
  • Complex pagination and query patterns that pushed past DynamoDB’s strengths.
  • PCI-sensitive data that needed low-latency encryption and solid operational controls.

What I built

  • The Spring Boot foundation for the greenfield service.
  • Stream processing over Kinesis for high-volume transaction ingestion.
  • A migration path from DynamoDB to PostgreSQL where richer query behavior mattered.
  • AWS KMS envelope encryption with key caching to cut latency.
  • Faster local and sandbox deployment loops with Docker-based automation and a local CDC setup.

Result

The platform shipped quickly, scaled to production traffic, and stayed workable for developers and on-call engineers instead of turning into an MVP that needed to be rebuilt later.