This document summarizes the benefits of building an in-house machine learning platform called Positron. Key points: - Positron allows for quick and consistent model deployments, simplified model management, experiment tracking, and efficient workflows. - It features a multi-model pipeline for seamless model creation and validation. Models can be deployed with minimal configuration. - The platform uses MLeap for model serialization/deserialization, which provides portability and fast performance without dependencies on specific frameworks. - It aims to provide low latency and high throughput predictions, while allowing for customization and integration with existing infrastructure. External and internal models can be easily deployed.