I’m thrilled to share that I successfully defended my Ph.D. in August 2025!
My dissertation, “AI-Enabled Framework for Robust Building Control Using Domain-Randomized Reinforcement Learning and Heuristic Policy Transfer,” explores how AI and reinforcement learning can enable scalable and uncertainty-aware building control.
I’m deeply grateful to my advisor, Prof. Panagiota Karava, and my committee members, Profs. Ilias Bilionis, Kevin J. Kircher, and Jianghai Hu, for their invaluable guidance and support throughout this journey.
I’m also excited to announce that our first paper from this work has just been published in Applied Energy!
You can freely access it here until December 02, 2025:
🔗 https://lnkd.in/gvWtvtTc
In this work, we developed a Model Universe–based control-oriented pipeline that uses domain randomization to generate synthetic yet physically consistent building models from limited construction information. Instead of requiring detailed calibration, the method samples plausible model variants using prior knowledge of building physics and typologies, enabling efficient model generation when design data are incomplete. By training RL controllers across this diverse model ensemble, the learned policies achieve robust and transferable performance under uncertainty, minimizing engineering effort during deployment.