Hi there, welcome to my profile! I work as a Postdoctoral Research Assistant at Queen Mary
University of London (QMUL). I obtained my PhD in Machine Learning and Materials
Informatics from QMUL in Septermber 2022. I have a keen interest in developing
new deep learning (DL) architectures to solve challenging problems in science.
I also enjoy solving data structures and algorithms
problems and front-end web development with HTML, CSS and JavaScript.
One of my research projects is particularly focussed on implementing graph neural networks (GNNs)
with integrated
attention mechanism to represent complex data types such as crystal structures and
chemical compositions for materials property prediction. See our journal article published in Advanced Science that proposed
a new GNN model with better performance compared to some other state-of-the-art ML models in
materials informatics. I also apply ML for the exploration of chemical composition space for new
materials discovery and validate ML results with ab initio calculations such as DFT.
Read my PhD thesis available online at Queen
Mary Library.
I frequently work with large quantities of structured data and apply data science
tools including python-pandas, matlab, SQL, etc., to process and capture interesting
patterns. I
taught Data Mining undergraduate module as a Teaching Fellow at QMUL and demonstrated several other
postgraduate modules
including Machine Learning and Artificial Intelligence. I have used web scraping and natural
language processing (NLP) techniques
to extract data from the internet for my research. I have some experience in using seq2seq models
for
language modelling.
I used to be a university rower and competed in four (M4-), pair (M2-) and single scull boats
as well as indoor rowing championships. I now play cricket and go cycling in my spare time. I enjoy
travelling and
good food, especially devilled chicken cooked with Sri Lankan spices.