Lita is a Hardware Engineer working in the AI and Advanced Architectures group at Microsoft. She received her B.S. degree from Caltech and her Ph.D. in Electrical Engineering at Stanford University. Her research interests are in hardware design for deep learning and AI algorithms. In particular, her work focuses on edge computing devices and bringing intelligence to mobile and embedded platforms using a combination of hardware-software co-design techniques. Her thesis explores the memory design implications for energy-efficient AI chips (using digital and mixed-signal circuit design), while exploiting the inherent error resilience of Convolutional Neural Networks for image classification tasks.
I am currently a Research Scientist at Reality Labs, Meta (formerly known as Facebook) working on hardware-software co-design for machine learning (ML) and artificial intelligence (AI) accelerators and energy-efficient edge computing devices for augmented reality (AR) & generative AI (GenAI) applications. My particular research interests are in cloud-to-edge computing, spanning from high performance cloud AI inference/servicing to on-device AI edge intelligence (tinyML and embedded AI).
Previously, I was a Senior Hardware Engineer working at Microsoft in the Azure Hardware Systems and Infrastructure group (formerly Cloud AI and Systems Technology, AI and Advanced Architectures) working on architecture and algorithm co-design for Transformer models (Generative AI, Copilot, Bing inferencing), LSTMs, and computer vision models. I completed my Ph.D. in Electrical Engineering at Stanford University in the Murmann Mixed-signal Circuits Group in December 2018. In particular, my thesis explored the memory design implications for energy-efficient AI chips (using digital and analog/mixed-signal circuit design), while exploiting the inherent error resilience of Convolutional Neural Networks for image classification tasks.
