A cloud-based platform that enables deployment and scaling of machine learning models on serverless GPUs with automatic resource management and low latency.

A feature that combines multiple inference requests server-side to increase throughput and optimize GPU utilization.