Hao Lu is a principal software engineer on the NVIDIA TensorRT-LLM team, specializing in large language model (LLM) inference. Before joining NVIDIA, Hao co-founded HippoML, a startup focused on generative AI inference. She also played a key role in the development of AITemplate and PyTorch Static Runtime, and contributed to open source projects including PyTorch, Caffe2, Caffe2Go, and QNNPACK during her time at Meta. Hao received her bachelor’s degree from Peking University and master’s degree and PhD from University of Notre Dame.