Machine learning and architecture
Solution architect Tellow
To provide a better user experience for customers and reduce costs I implemented a machine learning algorithm using Tensorflow. The neural network used classification on invoice scans to create bookings without the need for humans.
To further improve the user experience we wanted the machine learning service to process incoming data in realtime.
Realtime streaming architecture
The solution used several microservices communicating via binary event streams using the CQRS and event-sourcing patterns. Technologies used included Apache Kafka, Node.js, protocol buffers gRPC and GraphQL-subscriptions.