Jiu Xu
Engineering Manager, Reality Labs at Meta
California, United States
Summary
👤
Senior🎓
Top SchoolJiu Xu is an engineering leader with 11 years of experience building production AI and computer vision systems for AR/VR and autonomous driving, currently managing Reality Labs engineering at Meta in California. He has led delivery of consumer-scale features—real-time instant codec avatars and facial expression tracking powering >1M DAU—while previously architecting prediction stacks for Level 4 autonomy at Pony.ai and core hand-tracking pipelines for Apple Vision Pro. His background blends deep research (PhD from Waseda) and applied ML, including published work and patents on image search and super-resolution, and hands-on contributions to open-source PyTorch VDSR implementations. Known for translating novel CV research into reliable on-device pipelines, he combines product-focused delivery with a strong record in scalable systems and real-time ML.
11 years of coding experience
9 years of employment as a software developer
Bachelor’s Degree, Computer Software Engineering, Bachelor’s Degree, Computer Software Engineering at Nanjing UniversityDoctor of Philosophy (Ph.D.), Computer Science, Doctor of Philosophy (Ph.D.), Computer Science at Waseda University
English, Chinese, Japanese
Github Skills (14)
neural-network10
computer-vision10
pytorch10
deep-neural-networks10
python10
super-resolution10
modeling9
trainings9
evaluation9
eval9
machine-learning9
data-generation9
faster-rcnn8
mask-rcnn8
Programming languages (6)
Github contributions (5)
twtygqyy/pytorch-vdsr
Mar 2017 - Mar 2018
VDSR (CVPR2016) pytorch implementation
Role in this project:
ML Engineer
ML EngineerContributions:22 commits, 27 pushes, 1 branch in 1 year
Contributions summary:Jiu primarily contributes to the implementation of a VDSR (Very Deep Super-Resolution) model using PyTorch. Their work includes updating the model definition, incorporating gradient clipping, and integrating a pretrained model. Furthermore, the user added scripts for generating training data and evaluating the model's performance. They also introduced demo and evaluation scripts, showcasing the model's functionality.
pytorchdeep-neural-networkssuper-resolutioncvprvdsr-pytorch
twtygqyy/resnet-cifar10
Jun 2016 - Oct 2017
ResNet for Cifar10
Contributions:13 commits, 17 pushes, 1 branch in 1 year 4 months
resnetdeep-neural-networksresidual-networksdata-augmentationcifar10
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.