Lin Gong
Machine Learning Engineer at Meta
Sunnyvale, California, United States
Summary
🤩
Rockstar🎓
Top SchoolLin Gong is a machine learning engineer with nine years of experience applying statistical ML to real-world problems like query understanding, personalization, sentiment analysis, and user behavior modeling. Currently at Meta after leading query-intent work at Walmart Global Tech, Lin combines production ML engineering with research rigor from a PhD in Machine Learning and Data Mining. He has contributed to large open-source PyTorch projects, improving CI/CD for pytorch/builder and adding multimodal transformer features and attention caching to TorchMultimodal. Comfortable across model development, software engineering, and DevOps, Lin focuses on scalable pipelines and pragmatic model design that balance long- and short-term user signals. Based in Sunnyvale, he brings the discipline of academic research to high-impact product teams, often surfacing subtle engineering fixes that improve reliability in large-scale systems.
9 years of coding experience
4 years of employment as a software developer
Bachelor's degree, Information Engineering, Bachelor's degree, Information Engineering at Nanjing University of Aeronautics and AstronauticsDoctor of Philosophy (Ph.D.), Machine Learning and Data Mining, Doctor of Philosophy (Ph.D.), Machine Learning and Data Mining at University of Virginia
English, Chinese
Github Skills (20)
pytorch10
docker10
continuous-integration10
python10
scripting10
machine-learning10
bash10
dockers10
cicd10
transformer-models10
model-building10
script10
build-automation10
sh10
shell10
Programming languages (5)
Github contributions (5)
facebookresearch/multimodal
Mar 2022 - Oct 2022
TorchMultimodal is a PyTorch library for training state-of-the-art multimodal multi-task models at scale.
Role in this project:
ML Engineer
ML EngineerContributions:399 reviews, 250 commits, 127 PRs in 7 months
Contributions summary:Lin primarily focused on the development and improvement of models within the PyTorch multimodal library. Contributions included refactoring test targets, fixing mypy type errors in model code, changing test frameworks, implementing and updating position embedding APIs, making dropout optional in MLP layers, and adding caching to the MultiheadAttention mechanism. The user also added a checkpointing wrapper and implemented new features such as a multimodal transformer decoder and the VideoGPT model.
pytorchmulti-taskartdeep-learningmultimodal
pytorch/builder
Sep 2021 - Nov 2021
Continuous builder and binary build scripts for pytorch
Role in this project:DevOps Engineer
DevOps EngineerContributions:7 reviews, 18 commits, 10 PRs in 1 month
Contributions summary:Lin primarily focused on modifying and maintaining the CI/CD pipeline for the `pytorch/builder` repository. Their contributions involved updating configuration files, modifying build scripts, and addressing environment variable persistence issues across build steps. They also worked on reversing and reverting changes related to CI triggers and testing the directory structure within the CI environment. The user's changes focused on improving the automation of builds and integrating the builder with the continuous integration system.
pytorchbuild-scriptscontinuousbuilder
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.