Sihui Ding

As a Data Scientist at Autocase, I apply my passion for data science and energy efficiency to design deep learning models that forecast building energy usage. I hold a Master of Science in Operation Research from Columbia Engineering, where I gained the technical skills and knowledge to tackle complex data challenges and deliver innovative solutions. In my current role, I have improved the prediction model score significantly by creating a robust synthetic dataset, conducting thorough data validation, performing exploratory data analysis and data mining, researching and developing suitable neural network model to fit with the product requirement, and conducting iterative experimentation and evaluation. I have also reduced the model estimation time by 50% and achieved robust prediction results on a combined climate zone model, which helped reduce the future data extraction cost by 30%. Currently, I am working with XAI (explainable artificial intelligence) with SHAP and LIME, aiming to provide trustworthy model and helping others to understand and evaluate models, and OOD (out of distribution) detection, to build a robust model with great generalization ability. I am proficient in Machine Learning and Deep Learning Algorithms, Language Models, Python, and SQL. I am motivated by the opportunity to contribute to the domain of energy consumption forecasting with deep learning algorithm and to support the development of sustainable and cost-effective buildings.