Overview

Built a low-latency odds aggregation and pricing engine processing millions of market updates per day with sub-100ms p95 latency.

What I did

Designed a streaming pipeline with Kafka for provider feeds and Redis for hot caches.

Implemented a pricing microservice in TypeScript with circuit breakers and backpressure.

Wrote canary deployments and autoscaling policies to handle spikes during major events.

Highlights

Reduced odds propagation time from 450ms to ~85ms p95.

Cut cloud spend by ~18% via targeted caching and batch writes.