Interests and Hobbies

I've always been deeply interested in technology. From building an off-grid radio communications system, constructing neural networks to solve big data problems, and delivering a TED talk in NYC about renewable energy. More recently, however, my work has primarily revolved around the field of deep learning and AI. I am currently undertaking a 6-month internship as a data science intern at Xilinx, a U.S. semiconductor company, whilst also conducting research in the emerging field of geometric deep learning for wireless communication networks.

AI and Data Science

As data is becoming more prevalent in virtually all industries today, along with recent booms and breakthrough in the area of deep neural networks, I've taken it with great interest to pursue the field in greater depth. Some of my projects include:

• Using recurrent neural networks and gradient boosting to study the efficacy of medical drugs in treating schizophrenia
• Applying geometric deep learning (graph neural networks) to solve the power allocation problem in wireless communication networks
• Teaching a bot to play chess using the minmax and alpha-beta pruning algorithm
• Creating a chatbot to answer customer questions using deep learning in TensorFlow
• Used the NEAT (Neuro-Evolution of Augmenting Topologies) genetic neural network to solve a pathfinding problem.

The most sophisticated (and exciting) of these projects is studying the new field of geometric deep learning. Previously, deep learning models have been constrained to datasets with a direct Euclidean representation (i.e. data that can be represented in some multidimensional feature space). However, more recently, researchers have been investigating ways to train deep learning models on graphs and manifolds. Using graph neural networks (also known as message passing networks), I am studying its potential application in solving classical wireless network problems such as the power allocation problem more efficiently. Graph neural networks achieve this by performing an aggregation of node embeddings of neighbouring nodes over many layers.

Other projects I've undertaken is predictive modelling (e.g. classification with random forests and boosting), non-supervised learning with PCA and clustering, image recognition with convolutional neural networks, as well as basic natural language processing (NLP) with word embeddings and recurrent neural networks.

Hobbies

Outside of my studies, I really enjoy playing sports - including tennis, golf, running, swimming, and volleyball. I used to actively join local tennis competitions before university; though more recently, my main sport has been running and swimming. I'm also a huge fan of the olympics and the world cup whenever that's around. In addition, I also enjoy reading books every so often - my three most recent and favourite ones being A Brief History of Time by Stephen Hawking, Sapiens by Yuval Noah Harari, and Lord of the Rings by J. R. R. Tolkien.