Summary
Zhenghang Xu is a Senior Software Engineer in California with 13 years of experience building high-performance HW/SW co-design systems for GPUs and AI infrastructure. He has driven GPU system and emulation platforms, NVMe SSD controller design, and production deep learning inference stacks at NVIDIA and Amazon, blending low-level hardware insight with cloud-scale software delivery. At NVIDIA he now supports DGX, DGX Cloud, NeMo LLM features, and Triton/GPU Operator deployments while building RAG/chatbot solutions using LangChain and vector databases. Previously he led device inference frameworks and CV pipelines for Echo and Ring products at Amazon, automating model optimization to materially improve on-device performance. Known for connecting pre-silicon emulation and architecture verification to real-world inference performance, he brings a rare cross-domain fluency from ASIC layout to LLM fine-tuning. Trained in automation and electrical engineering (BIT, USC), he focuses on squeezing efficiency out of both silicon and software.
12 years of coding experience
11 years of employment as a software developer
Master of Science (M.S.), Electrical Engineering, Master of Science (M.S.), Electrical Engineering at University of Southern California
Bachelor of Engineering (B.E.), Automation, Bachelor of Engineering (B.E.), Automation at Beijing Institute of Technology
English, Chinese