Lei Mao
Systems Software Engineer at Meta
Palo Alto, California, United States
Summary
👤
Senior🎓
Top SchoolLei Mao is a systems software engineer with a decade of experience building deep learning systems and production-grade AI tooling, currently at Meta after multiple senior roles at NVIDIA. He blends a strong research background (MS in Computer Science from University of Chicago and an earlier MS in Biochemistry) with hands-on expertise in C++, CUDA, Python and ML model engineering. Lei has shipped and optimized large-scale deep learning infrastructure and contributed open-source projects in speech and computer vision—his CycleGAN-based voice conversion repo shows practical model architecture refinements and bug fixes in pitch conversion. Known for translating cutting-edge algorithms into readable code and user-friendly software, he also maintains technical blogs that make complex topics accessible. Based in Palo Alto, he brings both academic rigor and production-focused pragmatism to AI systems engineering.
10 years of coding experience
15 years of employment as a software developer
Nanodegree Machine Learning Engineer, Nanodegree Machine Learning Engineer at UdacityMaster of Science - MS Computer Science, Master of Science - MS Computer Science at University of ChicagoBachelor of Science - BS Biotechnology, Bachelor of Science - BS Biotechnology at Dalian University of TechnologyMaster of Science - MS Biochemistry, Master of Science - MS Biochemistry at Duke UniversityHigh School, High School at Ningbo Xiaoshi High School
English, Chinese
Stackoverflow
1reputation
64reached
0answers
1question
Github Skills (10)
machine-learning10
voice-conversion10
tensorflow10
speech-processing10
architectures10
cyclegan10
architect10
architecture10
python9
lambda6
Programming languages (21)
Github contributions (5)
leimao/Voice-Converter-CycleGAN
May 2018 - Oct 2019
Voice Converter Using CycleGAN and Non-Parallel Data
Role in this project:
ML Engineer
ML EngineerContributions:34 commits, 19 pushes, 52 comments in 1 year 4 months
Contributions summary:Lei's commits primarily involve modifications to the `model.py` and `module.py` files, indicating a focus on the core implementation of a CycleGAN for voice conversion. These changes include updates to the model architecture, such as incorporating gated CNN layers, adjusting loss functions, and refining the training process with learning rate decay and validation steps. Furthermore, the user addresses a bug in the pitch conversion mechanism, suggesting a hands-on role in model refinement and debugging.
cycleganparallelparallel-datavoiceconverter
leimao/PyTorch-Static-Quantization
Nov 2020 - Apr 2021
PyTorch Static Quantization Example
Contributions:15 commits, 2 PRs, 5 pushes in 5 months
pytorchquantizationmachine-learning
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.