Ramp's initial product growth was driven by a set of opinionated best practices, but as complexity increased, more flexible solutions were needed. The company transitioned to a unified platform by building a generic workflows engine, improving flexibility and scalability. The solution involved using a graph-based data structure to enable complex workflows and integrating it with asynchronous task management backed by Postgres. The approach allowed for easier feature creation and significant performance improvements. The platform now supports various product functionalities and can handle over a million workflows daily.
Sort: