Xu Dong
Software Engineer at Meta
Menlo Park, California, United States
Summary
👤
Senior🎓
Top SchoolXu Dong is a software engineer based in Menlo Park with 11 years of experience applying machine learning and computer vision to products at scale, currently building Search AI at Meta. His background spans eye tracking for Oculus, autonomous driving research (stereo depth and optical flow) at TuSimple, and search ranking systems, reflecting a blend of applied research and production engineering. He has contributed notable operators and ConvLSTM/ConvRNN implementations to the widely used Apache MXNet project, evidencing deep expertise in CNNs and sequence modeling. Comfortable moving models from research into robust systems, he focuses on image-intensive modalities and cross-sensor calibration problems that bridge perception and product needs.
11 years of coding experience
6 years of employment as a software developer
Master Electrical and Computer Engineering, Master Electrical and Computer Engineering at Purdue University - Office of the Vice Provost for Graduate Students and Postdoctoral Scholars
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (10)
cuda10
mask-rcnn10
faster-rcnn10
mxnet10
machine-learning10
rnn-model10
deep-learning10
n10
python10
image-processing9
Programming languages (2)
Github contributions (5)
apache/mxnet
Jul 2016 - Aug 2017
Lightweight, Portable, Flexible Distributed/Mobile Deep Learning with Dynamic, Mutation-aware Dataflow Dep Scheduler; for Python, R, Julia, Scala, Go, Javascript and more
Role in this project:
ML Engineer
ML EngineerContributions:7 commits, 14 PRs, 23 comments in 1 year 1 month
Contributions summary:Xu contributed significantly to the implementation and documentation of operators within the MXNet framework, primarily focusing on deep learning operations. Their work includes adding new operators like "correlation" and "BilinearSamplingOp" along with related unit tests, demonstrating expertise in convolutional neural networks (CNNs) and image processing. Furthermore, the user implemented convolutional RNN cells such as ConvLSTMCell, showcasing advanced knowledge of recurrent neural networks (RNNs) and their application in sequence modeling. The user updated documentation of core operations.
pythonschedulerdataflowmutationdata-science
dsqx71/mxnet
Jul 2016 - Aug 2017
Contributions:52 pushes, 1 branch in 1 year 1 month
pythonschedulerdataflowmutationdata-science
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.