Zihan Zhang

Zihan Zhang

Ph.D student at University of Technology Sydney

About Me

👋 My name is Zihan Zhang (in Chinese, 张梓涵). I’m a third-year Industry Doctorate Program (IDP) PhD student at the University of Technology Sydney (UTS) NLP group, sponsored by the industry partner TPG Telecom. I am privileged to be supervised by Prof. Ling Chen, Dr Mo Namazi, and Distinguished Prof. Jie Lu. I am also working closely with A/Prof. Meng Fang from the University of Liverpool. I worked at TPG Telecom as a Research Analyst in the Data and Analytics Centre of Excellence (CoE) team for about three years. Prior to joining UTS, I obtained my Master degree in Software Engineering from the University of Melbourne.

I have broad research interests centred around natural language processing (NLP). As an IDP PhD student, I am particularly passionate about developing NLP systems that are practical, robust, efficient, and can be applied to the industry. While working at TPG Telecom, I had the chance to apply my skills in solving real-world business challenges involving millions of telco data. Specifically, I developed:

  • Clustering-based topic modelling to help the Call Centre team find top topics from millions of customer feedback/comments/survey (NPS) to improve service;
  • An efficient data annotation method for labelling raw dialogue state data, which is essential for building AI-powered chatbots;
  • A Market Offer Engine to automatically collect hot offers from the market, which provides an analysis data source for the product team to make pricing strategies;

Recently, my research focus has moved towards Large Language Models (LLMs) and exploring how to use them in production to drive business. My recent efforts involve adapting LLMs to world knowledge, which includes developing efficient, trustworthy, and robust Retrieval-Augmented Generation (RAG) methods for Question Answering (QA) and Dialogue systems.

While studying Software Engineering, I have gained skills in designing, modelling, and building software applications, especially web applications. Check out previous 🔨 software applications I built with my teammates if you are interested.