Haitong Li
Assistant Professor at Purdue University
Palo Alto, California, United States
Summary
🤩
Rockstar🎓
Top SchoolHaitong Li is an Assistant Professor of Electrical and Computer Engineering at Purdue University with a decade of experience bridging device physics, circuits, and ML-aware hardware for next-generation computing. Trained at Peking University and Stanford (PhD EE, 2022), he has blended academic rigor with industry impact through research roles at Stanford, Meta/Reality Labs, Facebook Reality Labs, and Arm. His work focuses on new hardware technologies across the stack—particularly non-volatile memories and neural accelerators—aimed at enabling ubiquitous machine intelligence. He also contributes practical ML tooling, for example a PyTorch knowledge-distillation framework that supports teacher-student experiments and experiment logging. Based in Palo Alto, he recruits students to translate compact device models and circuit techniques into system-level hardware innovations. Colleagues note he combines deep experimental device expertise with hands-on ML engineering, enabling cross-disciplinary projects that go from materials to models.
10 years of coding experience
6 years of employment as a software developer
Bachelor’s Degree, Microelectronics, Bachelor’s Degree, Microelectronics at Peking UniversityMaster of Science - MS, Electrical Engineering, Master of Science - MS, Electrical Engineering at Stanford University
Github Skills (10)
neural-network10
pytorch10
distill10
deep-neural-networks10
dis10
model-compression10
computer-vision9
faster-rcnn8
mask-rcnn8
tensorboard6
Programming languages (3)
Github contributions (5)
haitongli/knowledge-distillation-pytorch
Mar 2018 - Sep 2022
A PyTorch implementation for exploring deep and shallow knowledge distillation (KD) experiments with flexibility
Role in this project:
ML Engineer
ML EngineerContributions:129 commits, 5 PRs, 124 pushes in 4 years 6 months
Contributions summary:Haitong implemented and refined a PyTorch-based knowledge distillation framework. Contributions include adding the necessary training functions for knowledge distillation (KD) and its evaluation. They added the KD loss function, modified the training and evaluation pipelines to incorporate the teacher model, and introduced the ability to load pre-trained teacher models to enable various KD experiments. Furthermore, the user enhanced the training process, including the learning rate scheduler and the addition of TensorBoard logging.
pytorchdistillationknowledge-distillationknowledgedeep-learning
peterliht/peterliht.github.io
Jul 2018 - Aug 2022
Contributions:29 pushes, 1 branch in 4 years 1 month
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.