I am a 5th year Master’s student in Computer Science at Carnegie Mellon University, advised by Professor Chenyan Xiong at the CMU Foundation and Language Model Center (FLAME). I completed my Bachelor’s degree in Computer Science at CMU in 2025, where I developed a strong foundation in computer systems and machine learning.
My research focuses on model-infra co-design for scalable and efficient foundation models, with particular emphasis on architecture innovations such as Mixture of Experts (MoE) and natively trainable sparse attention mechanisms that enable conditional computation and long-context reasoning at scale. I am passionate about bridging the gap between theoretical advances and practical systems, building research platforms that accelerate scientific discovery while maintaining production-level engineering standards.
During my undergraduate years, I was the main contributor to multiple research projects spanning large language models and information retrieval. My work has appeared at EMNLP and NAACL, including research on interpretable dense retrieval, autonomous research agents, and LLM-generated text detection. Most recently, I built FLAME-MoE, an open end-to-end research platform for Mixture-of-Experts language models that provides transparent access to all training artifacts, enabling the community to study expert specialization and routing dynamics at scale.
Beyond research, I have hands-on experience building production systems during my internship at Glober AI, where I developed scalable video inpainting services optimized with TensorRT. I enjoy tackling challenging systems problems, from implementing Unix-like operating system kernels to building distributed deep learning frameworks with CUDA and NCCL from scratch. In the end, I love writing code that makes things happen!
Feel free to reach out if you’re interested in discussing research, collaboration opportunities, or anything related to foundation models and systems.