Biography
I am a Senior Machine Learning Scientist at TikTok. I received my Ph.D. in Oct 2024 from UCLA, where I was honored as a recipient of the Amazon Fellowship and the UCLA Dissertation Year Fellowship.
Currently, I architect industry-scale Multi-modal LLMs (MLLM) and generative recommendation engines. My research focuses on bridging advanced deep learning with physical principles, including self-supervised learning, long-sequence modeling, and computational imaging.
Latest News
January 2026
"Multi-scale Conditional Generative Modeling for Microscopic Image Restoration" accepted by IEEE ICASSP 2026.
Selected Works
Hallucination detection in virtual tissue staining and digital pathology
Introduced a monitoring framework using cycle-consistency and uncertainty quantification to identify neural network hallucinations, ensuring high-fidelity results in AI-driven pathology.
Read PaperSelf-supervised learning of hologram reconstruction using physics consistency
Developed GedankenNet, a self-supervised model eliminating the need for labeled training data through physics-informed consistency and artificial data.
Read PaperFourier Imager Network (FIN): Superior External Generalization
A Fourier-based framework processing spatial frequencies with learnable filters, exhibiting superior generalization and inference speed compared to standard CNNs.
Read PaperRecent Publications
Education
Ph.D. in ECE
UCLA | 2019 – 2024
B.Eng. in Optical Science
Zhejiang University | 2015 – 2019