Skip to main content

Components

The Fraugster Engine is composed of a set of distributed services which are all written in Go and communicate via gRPC.

We use a variety of Open Source technologies:

  • Etcd as a key-value store,
  • Kafka as a message bus,
  • Athena as a primary data storage,
  • TensorFlow to run neural networks,
  • Elasticsearch to power search.

All components run in a fully containerized environment on Kubernetes. Our core scoring component is based on an in-house developed in-memory database which provides high performance query operations for the scoring engine.