Case Study: Melio scales risk decisioning and feature development with Chalk

Melio opens feature development to 20+ contributors with Chalk

Melio, a B2B payments fintech processing billions of dollars for small businesses, faced significant scaling challenges with its homegrown Python feature store. As payment volume grew, the in-house system became a bottleneck, absorbing engineering bandwidth and limiting feature development to only a handful of engineers. This infrastructure could not reliably support the demanding needs of their risk decisioning for fraud, compliance, and non-sufficient funds. To address this, Melio chose to replace their internal system with Chalk, a managed feature platform.

By implementing Chalk, Melio replaced its entire in-house feature store. The solution provided mission-critical uptime, connected to all of Melio's data sources, and enabled self-serve feature development. This shift freed engineers to focus on model building and allowed over 20 developers and data scientists across the risk organization to independently write and deploy features. The results included a fully managed feature store with zero internal headcount, risk workloads running separately from production, and expanded use cases into product analytics, greatly improving stability and scalability.

Melio