I am a researcher in the School of Mechanical Engineering at Tongji University, specializing in Heating, Ventilation and Air Conditioning (HVAC). I am currently a visiting scholar at the Center for the Built Environment (CBE), UC Berkeley.
My research focuses on energy forecasting, optimal operation and control for HVAC systems in commercial buildings using data science technologies. Specifically, I leverage causal science and causal machine learning to enhance the physical principles underlying these models in the context of building energy data. I am also interested in automating energy management and optimization tasks with large language models (LLM), primarily focusing on automating energy efficiency diagnosis using multi-source building data.
📰 News
🎓 Education
-
2019.09 - 2025.12Ph.D. in Mechanical EngineeringTongji University. GPA 4.93/5.0. Supervised by Prof. Peng Xu.
-
2015.09 - 2019.06B.E. in HVAC EngineeringTongji University. GPA 4.61/5.0.
💼 Work Experience
-
2024.12 - NowVisiting Scholar
-
2023.07 - 2023.08Research Intern (Applied Energy Trainee Program)DC Building Lab, SZIBR. Developed control strategies combining building flexibility and occupant behavior.
📝 Selected Publications
-
Exploring automated energy optimization with unstructured building data: A multi-agent based framework leveraging large language models Q1, IF=6.6Energy and Buildings, 2024 -
An interpretable method for identifying mislabeled commercial building based on temporal feature extraction and ensemble classifier Q1, IF=11.7Sustainable Cities and Society, 2022 -
Status quo and opportunities for building energy prediction in limited data Context—Overview from a competition Q1, IF=11.2Applied Energy, 2022
🚀 Project Gallery
💻 GitHub Projects
A specialized Question-Answering (QA) LLM chatbot tailored for the HVAC domain. Includes entire frontend, backend, and database. This is a personal project developed in 2023.
A robust, rule-based algorithmic engine for computing building energy consumption baselines and carbon emission limits. This is a personal project developed in 2022.
My solution (8/~1000) for the 'Causal Learning and Decision Optimization Challenge' of the World Artificial Intelligence Conference (WAIC) 2022 Hackathon.
An expandable machine learning framework practice project created in 2022. Provides a highly extensible, object-oriented architecture to standardize the pipeline of building, tuning, and evaluating any machine-learning-based energy model.
A Datawhale China open-source tutorial. Based on 'Design Patterns' by Michael C. Feathers, this project interprets design principles, design patterns, and provides coding examples.
A Python package for converting heating hot water system data to Brick Schema models with comprehensive validation and portable analytics.
🏆 Selected Honors & Awards
- 2022 - Outstanding Doctoral Scholarship, Tongji University.
- 2022 - Outstanding Student, Tongji University.
- 2022 - Datawhale Contributor.
- 2019 - Outstanding Graduate, Shanghai.
- 2017 - Second Prize of National Mathematical Contest in Modeling, China Society of Industrial and Applied Mathematics.
💬 Selected Presentations
-
2025Toward Automated Building Performance Analysis with LLM-Based AgentsBuildNext: A Global Seminar for Young Researchers. @Syracuse, NY, USA.
-
2025Causal analysis of distribution shift in building energy modelsWeLL Seminar Spring Student Spotlight Talk. @Berkeley, CA, USA. [Picture]
![[Picture]](https://sheltonxiao.github.io/images/files/WeLL2024.jpg){kind=link}