Real-Time Machine Learning And AI For Medical Billing

How MiddleGate is Using Akka to Disrupt the $471 Billion Medical Billing Industry

Yet another dynamic to the ongoing challenges for the U.S. healthcare system is the $471 billion annually that it spends on billing and insurance-related (BIR) paperwork1. $375 billion of that is estimated to be wasteful.

Portland, Oregon- based MiddleGate is targeting the claim process between healthcare providers and insurance companies--where underpayments and denials cost healthcare providers billions. It’s a huge opportunity for technology disruption, and MiddleGate is a hot startup that’s already serving medical billing companies supporting more than 1,000 hospitals and doctors offices.

We were excited to connect with Middlegate’s Roshan Fernando (CEO) and Ed Stull (CTO) to learn more about their use of Akka as the backbone of their Grace platform, and how their new approach is taming key medical billing complexity issues that apply to every healthcare provider in the United States:

Listen on Soundcloud


  • How the Akka model is foundational to MiddleGate’s Machine Learning processes for bringing more intelligence to claims resolution, by creating analytics processes as Actors that can continuously learn from inputs and feed the Grace Machine learning model, deliver outputs against payment claims, and operate across organizational boundaries.
  • How Akka and Actors allowed MiddleGate to create an distributed system that allows both isolation at the local level (clients on their own computing resources on-prem), and sharing at the global level (cloud-based).
  • How MiddleGate used Akka to simplify cloud-based deployment of its platform to medical billing companies and their networks of clients (hospitals and doctors offices)--making it possible to scale the business in a way that surpassed previous methods.
  • How Akka plugged in to key MiddleGate requirements around isolation, concurrency, immutability--and how it was through Akka that MiddleGate overcame the scale limitations of the first generation (database and SQL procedures + Java EE) and second generation (MQ + Vertx) systems.


Visit for an overview of our subscription offerings or download the Lightbend Subscription data sheet for detailed information