Zhengfeng Lai
Senior Research Scientist Tech Lead, World Model at MBZUAI Institute of Foundation Models
Cupertino, California, United States
Summary
🤩
Rockstar🎓
Top SchoolZhengfeng Lai is a senior research scientist and tech lead specializing in world models, multimodal and video large language models, and data-centric approaches to image/video generation with nine years of research and industry experience. He has led VLM training efforts from compact architectures to extremely large Mixture-of-Experts models and co-developed open-source world models, including the PAN world simulation release. At Apple and MBZUAI he drove video foundation-model research and practical systems for long-horizon, token-efficient video understanding and streaming video assistants, and his work spans few-shot, semi-supervised and unsupervised domain adaptation methods. His background includes applied roles at Amazon Lab126 and entrepreneurial leadership in medical-device ML, where he translated user research into a neonatal screening product. He holds a PhD from UC Davis and has EB-1A approval, reflecting recognized impact in AI research. Based in Cupertino, he combines deep academic rigor with product-minded engineering and a persistent focus on label- and data-efficiency in real-world multimodal systems.
10 years of coding experience
6 years of employment as a software developer
Doctor of Philosophy - PhD, Electrical and Computer Engineering, Doctor of Philosophy - PhD, Electrical and Computer Engineering at University of California, DavisBay-Area Regional Innovation CORPS Training, Bay-Area Regional Innovation CORPS Training at University of California, Berkeley, Haas School of BusinessBachelor of Engineering - BE, Information Engineering, Bachelor of Engineering - BE, Information Engineering at Zhejiang University
Github Skills (41)
speaker-diarization10
representation-learning10
pytorch10
machine-translation10
artificial-intelligence10
speaker-identification10
residual-networks10
speech-processing10
python10
speaker-verification10
chainer10
speech-enhancement10
text-to-speech10
apc10
speech-synthesis10
Programming languages (4)
Github contributions (5)
jefflai108/VGNSL
Dec 2021 - Feb 2023
[ACL 2019] Visually Grounded Neural Syntax Acquisition
Contributions:146 pushes, 1 branch in 1 year 2 months
aclacquisition
jefflai108/ASSERT
Apr 2019 - Oct 2019
JHU's system submission to the ASVspoof 2019 Challenge: Anti-Spoofing with Squeeze-Excitation and Residual neTworks (ASSERT).
Contributions:26 commits, 32 pushes, 1 branch in 7 months
residual-networksasvspoofanti-spoofingspoofing-attackautomatic-speaker-verification
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.