PhD researcher in robust AI, generative AI, and accelerated learning

Rigorous AI for generative systems, defect detection, and accelerated learning.

I build mathematically grounded AI across generative models, robust defect detection, graph learning, stability analysis, diffusion methods, and training acceleration for industrial and research systems.

Recent highlights include a DAC 2025 Best Paper Nominee, an SPIE Advanced Lithography + Patterning paper, two KLA invention disclosures, and scientific advisory work across startup and industrial AI programs.

Open CV Browse Projects Contact
CV updated Loading...
Projects updated Loading...
Portrait of Wuxinlin Cheng
Research approach Math and theorem-driven, not heuristic-first

I build from spectral theory, probabilistic graphical models, and stability analysis to explain why models work and when they fail.

Industry and advising KLA AI engineer and scientific advisor across applied AI teams

Delivered 40% and 25% gains on ID and OOD metrology tasks, 5.8x runtime improvements, and production robustness wins in industrial vision.

Research record ICML, CVPR Demo, DAC, SPIE, ICLR reviewer, 2 disclosures

Published across ML, vision, EDA, and semiconductor venues while also producing internal invention disclosures for industrial AI.

Selected work

Loading projects...