👋 Hi! I am Shih-Yu Lai, a Master’s student at National Taiwan University, where I am a member of the Communications and Multimedia Laboratory and Interactive Graphics Laboratory, advised by Prof. Robin Bing-Yu Chen. My research focuses on sim-real learning and computational solvers for physical and embodied systems, spanning artificial intelligence, computer graphics, robotics, and audio. I work with physics-informed machine learning, neural PDE/ODE models, reinforcement and robot learning, graph learning, and generative/multimodal models.
My research experience spans international academia and industry. In 2026, I joined Aalto University through the Aalto Science Institute Summer Research Programme, working with Prof. Vikas Garg and the QuML Group on machine learning and computational methods. At Aalto, I also collaborated with researchers from the Aalto Acoustics Lab, including Nils Meyer-Kahlen and Eloi Moliner, on audio, spatial acoustics, and learning-based signal processing.
Previously, I was an HPC Research Intern at the RIKEN Center for Computational Science (R-CCS), working with Prof. Hirozumi Yamaguchi on large-scale digital twins, multimodal WiFi/point-cloud sensing, federated learning, and high-performance computing on Fugaku. I later joined MoonShine Animation Studio as a VLM Research Intern, developing multimodal AI tools for animation and VFX production, including creator support for storyboarding and post-production. Earlier, as a Deep Learning Intern at Delta Electronics, I developed point-cloud representations and predictive models integrating finite-element materials analysis with electromagnetic simulation.
Before NTU, I studied at National Cheng Kung University, where I worked with the Robot-Aided Creation and Construction Research Group led by Prof. Yang-Ting Shen and Prof. Chia-Ching Yen on computational assembly, graph reinforcement learning, robotic fabrication, and CAD-based design. I was also affiliated with the Urban Science and Computing Lab led by Prof. Hsun-Ping Hsieh, where I worked on urban digital twins and 2D/3D spatial modeling.
💻 I primarily work in Python, with experience in C++, ROS/ROS2, OpenGL.
