Scalable ML Serving Architecture

A horizontally-scalable inference service with batching, caching, and graceful degradation.

Designed a system that serves model predictions under load with request batching, a Redis cache layer, and circuit breakers.