A lifelong learner, Binnie received a BEng in Mechanical Engineering at Hong Kong Polytechnic University in , then worked as an acoustic consultant for six years. Unsure this was right career direction and always keen on mathematical theory and applications, he then undertook an MSc in Mathematics at the Chinese University of Hong Kong. He joined CLP in as a procurement o cer, continuing to broaden his knowledge to arti cial intelligence through self-learning and online research, ahead of applying for the dual master’s degree program. In devising the model for his MPhil thesis, Binnie was supervised and inspired by Prof. ZHANG Tong, Chair Professor of Mathematics and Computer Science & Engineering at HKUST as well as CLP industry supervisor Dr. LEE Cheuk-Wing. Meanwhile, he gained additional practical insights from his online distance learning degree courses at Strathclyde, which is renowned for its industry-based expertise in electricity and power. These included power system fundamentals and renewable energy, among others. In his thesis, Binnie advocated combining Prof. Zhang’s Regularized Greedy Forest (RGF) algorithm, which employs a tree-based machine learning model to form the decision forest, together with two gradient boosting models (XGBoost and LightGBM), creating an ensemble model that delivered improved stability and accuracy. On the application front, during his studies he was transferred to work as an engineer in the condition monitoring team in CLP’s Technical Services Department, using temperature and heat indexes from three Hong Kong Observatory weather stations as input for his model. In , together with CLP’s System Operation Department, Binnie started to develop, test, and implement the ensemble model. By early , they had successfully introduced the model to CLP’s daily operations. It is now producing forecasts that enable system operators to e ectively manage demand responses, optimize generation resources, and simplify and facilitate operational planning decisions. Further bene ts include enhancement of equipment servicing decision-making while ensuring su cient energy supply to customers, he noted. On the sustainable development front, better system forecasting means fuel is saved, assisting decarbonization. Binnie also hopes that the model’s use can eventually be extended from system operations to circuit and consumer levels, widening its reach and encouraging further smart energy use. To complete his dual degrees in less than three years while working full-time, Binnie balanced his time by setting priorities while tting sports activities in between to release stress and keep himself on an even keel. His motivation comes from his belief that you should continuously seek to improve yourself and do your best – and from his engineer father, whom he considers to be a role model. Looking to the future, he would like to keep focusing on both his career and research on machine learning, nding his greatest rewards in creating things that work well. “For me, it’s not about studying for studying’s sake. I am also committed to industry. By combining the theoretical and practical, that’s how we can really make an impact.” Given this outlook, it appears Binnie will soon be adding to his collection of rsts. * Graduates will be awarded a HKUST MPhil degree in one of these disciplines depending on their academic pathway and thesis project: Chemical and Biomolecular Engineering, Civil Engineering, Computer Science and Engineering, Electronic and Computer Engineering, or Mechanical Engineering. 25 IN FOCUS Alumnus Binnie Yiu at HKUST, and (below, left) with the University’s former Dean of Engineering Prof. Tim Cheng, now Vice-President for Research and Development, and (below, right) Strathclyde University Principal and Vice-Chancellor Prof. Sir Jim McDonald.