Healthcare API: 60s+ → ~7s

03 decisions

1. 01

   Replace the HTTP integration with the OystEHR SDK

   Replaced the HTTP-based backend integration with the OystEHR SDK, for reliability.

2. 02

   Aggregate requests

   Optimised the NestJS APIs with request aggregation, so one call does the work of several.

3. 03

   Tune the database

   Optimised PostgreSQL queries as part of the same pass. Together with the aggregation layer, response times fell from 60s+ to about 7s.

4. 04

   Modernise on AWS

   Part of a wider modernisation of the platform on AWS.

$ ls ./case-studies

• Monolith → microservices on Google CloudLed the migration of a confidential enterprise SaaS platform from a monolith to event-driven microservices — downtime −40%, deployment time −50%.
• GraphQL performance layer: +65%Improved an enterprise app's performance by 65% with GraphQL query batching and caching, and delivered a Firebase Cloud Messaging notification pipeline end to end.
• Customer health platform for a VoIP providerDesigning an internal platform for Primo Dialler, a telecoms/VoIP provider, that brings call data and carrier traffic into a single view of each customer's health.
• PowerPick — a fantasy cricket league appBuilt end to end for Bitsol: a cross-platform React Native app that replaced a manual Google Forms and Excel workflow, on a fully serverless Firebase backend.
• OpenTask — a project-management platformMy final-year project: a complete MERN project-management platform with task and milestone tracking, team coordination, and real-time chat and notifications over WebSockets.
• AI integration & frontend systemsBuilt React/TypeScript interfaces for multi-agent AI systems, integrated OpenAI to cut manual email drafting by 70%, and delivered end-to-end Firebase Cloud Messaging pipelines.