Dr Sevvandi Kandanaarachch

Sevvandi is an interdisciplinary researcher working in data science and analytics at RMIT University.  She uses mathematics, statistics, and machine learning for her research. Her main research interest is anomaly and event detection. Examples of anomalies include credit card fraud, intrusions in computer networks, malfunctioning sensors, stock market crashes and astronomical anomalies such as solar flares. It is important to identify these anomalies quickly because it enables us to act upon them. 

Sevvandi has experience working on real-world applications. From 2016 to 2019, she worked with an industry partner on intrusion detection as part of an ARC Linkage project. She also worked with QUT researchers and the QLD Government on a water quality project as part of an ACEMS (ARC Centre for Excellence for Mathematical and Statistical Frontiers) research collaboration, which was featured by the ARC.  She has also worked on DELWP’s Vegetation Detection Challenge, which addressed aspects of bushfire mitigation. 

On the data frontier, Sevvandi has experience working with complex and high-volume data. She has worked with sensor data from satellites, water quality sensors and pedestrian count sensors.  Generally, sensors provide a non-stop data stream. Streaming data provides many challenges including high dimensionality, non-stationarity and partial information. Efficient algorithms are needed for processing such data to gain meaningful insights.