This is a unforgettable data science journey, and I am excited that I have been interviewed by the Hong Kong University Business School to share my experiences in leading a team to complete the USDCNH trading capstone project. I am grateful to the support from bank, my teammates and the university, provided an immersive opportunity to analyze the technical and economic factors influencing trading using machine learning, which was not just merely an academic endeavour but also solving real-world financial challenges with the applications of mathematics and statistics in the realm of data science. With this experience from University, I joined Google’s machine learning group, and opened doors to an advanced exploration of machine learning, taking a step further into the realms of deep learning and statistical modelling. Exposure to Google’s cutting-edge practices and the collaborative environment fueled my enthusiasm for pushing the boundaries of what is possible with machine learning.

This transformative year of intellectual growth, ignited by my first Data Science Program lecture at HKU Main Campus last April, has redefined my understanding of modern treasury management. Embarking on this dual frontier of machine learning and financial engineering, I’ve systematically decoded Python’s pandas library for cash flow forecasting, mastered machine learning volatility pattern recognition, and explored deep learning’s potential in algo-trading – each skill hard-won through midnight coding sessions and countless iterations.

The true challenge lay not in absorbing technical knowledge, but in bridging two alien worlds: financial rigor meeting computational creativity. Early attempts to apply clustering algorithms to liquidity pools revealed un-structured transactional and trading data required normalization, while timestamp inconsistencies. This forced me to confront the brutal reality of real-world data science: before any elegant model comes the gritty trench warfare of data cleansing. After hard work, the data can be used for further machine learning and fit the technical analysis data for learning. The prediction for currency market is more reliable and has been enlightening, I am poised to leverage these experiences to contribute meaningfully to the transformative landscape of fintech in the future