Serverless ML Inference with AWS Lambda + Docker

Running ML models in production sounds simple until you realize you're paying for servers 24/7 even...