I’m currently working as a researcher at Nokia Bell Labs in Antwerp. Our team is developing World Wide Streams, a platform for real-time stream processing.

The emergence of the “Internet of Things” has lead to a proliferation of sensors and actuators that are constantly exchanging streams of data. These streams range from small sensor samples, such as reading a speedometer in a car, to high-bandwidth video streams, e.g. a dash-cam. Then, applications combine streams and continuously process their data to detect notable events and act upon them, e.g. detecting when a car might run into another and brake automatically.

Our team at Nokia Bell Labs is developing World Wide Streams (WWS): a platform on which such applications for real-time stream processing can be developed. WWS can ingest and process data and video streams in many different formats. An application built on WWS consists of a dataflow, which is composed of several operators. These operators can be one of a set of built-in operators (ranging from map and filter to object detection) or custom-made by the developer. The operators are then “wired up” in a language called XStream, a library for TypeScript.

My current focus is on how to make programming such massively distributed, massively parallel applications easier – by developing suitable programming languages and techniques.