In this project, I designed and implemented the backend of an e-commerce platform and improved its performance to handle high-volume traffic (to 10000 queries per second).
- Performed ETL with Java to seed the database with 25 million records in 40 min.
- Optimized local average query time to <30 ms through indexing and token-based pagination
- Containerized the server and database with Docker and deployed them on AWS EC2.
- Conducted load tests with Loader.io and New Relic to identify performance bottlenecks.
- Scaled out the backend with an HAProxy load balancer and six servers, achieving a throughput increase from 1000 to 10000 QPS with 70 ms average latency and 0.3% error rate.