I am a high school student interested in artificial intelligence, efficient deep learning, and large language models. My research focuses on understanding the optimization and routing behavior of sparse neural networks, particularly Mixture-of-Experts (MoE) architectures. As of competition experiences, I won Bronze on USAAIO, Silver on USACO, and is a finalist on IAI2O (2025-2026).
I am currently working on my first research project investigating how pretrained attention can guide expert routing in sparse Mixture-of-Experts Vision Transformers. My work studies the relationship between routing diversity, optimization stability, and expert specialization through controlled experimental analysis. [Details will be included after publish]
Mixture-of-Experts models · Vision transformers · Sparse Neural Networks · Deep Learning · Competitive programming (USACO, USAAIO)
GitHub · AttnRoute-MoE Repository (Will publish after paper publication) · ORCID · Current High School Email
Last updated: July 2026