About

Who I Am

I am an AI Infra Engineer at Infinigence-AI (无问芯穹), working on efficient LLM inference. Before Infinigence-AI, I was at DeepSight, Cambricon, and UISEE — and before that, @UESTC and @HEU.

My main job is the LLM inference engine: LLM & AIGC inference, kernel optimization, agentic infra, and everything that makes a model actually run fast on real hardware. Along the way I have worked on computer vision — feature matching, morphology, and camera calibration.

Contact

• Blog: CSDN
• Zhihu: duanyzhi
• Email: duanyzhi@outlook.com