Ming Huang

Ming-Xu (Ming) Huang is a senior developer technology engineer at NVIDIA. He specializes in optimizing LLM training workflows, driving peak performance across the entire compute stack from end-to-end pipelines down to kernel-level.
Avatar photo

Posts by Ming Huang

Agentic AI / Generative AI Jul 10, 2026

Reducing High-Bandwidth Memory Bottlenecks in JAX-Based LLM Training with Host Offloading

Large language model (LLM) training workloads increasingly run into GPU memory limits before compute is fully used. Model weights, gradients, optimizer states,... 9 MIN READ