Fei Gao - Software Engineer at ByteDance

Fei Gao

Software Engineer at ByteDance

Sunnyvale, California, United States

Join Prog.AI to see contacts

Summary

🤩

Rockstar

🎓

Top School

Fei Gao is a software engineer with a decade of experience building large-scale distributed machine learning systems and production ranking infrastructure, currently working on TikTok ads ranking at ByteDance in Sunnyvale. Previously a researcher at Microsoft Research Asia, Fei contributed core back-end and infrastructure code to high-profile open-source projects such as Microsoft’s Multiverso parameter server, LightLDA, and CNTK, enabling efficient distributed training and scalable topic modeling. Their work spans low-level server components, inter-process communication (ZeroMQ), data preprocessing, and integration of ASGD for distributed deep learning, reflecting strong systems-level engineering for ML workloads. Fei holds a Master’s in Machine Learning from Beihang University and brings both research rigor and production-focused delivery to cross-functional teams. A notable, less obvious strength is the combination of research-era contributions to foundational ML tooling and current product-facing experience in real-time ad ranking, making them adept at moving algorithms into high-throughput production environments.

10 years of coding experience

4 years of employment as a software developer

Master, Machine Learning, Master, Machine Learning at Beihang University

Github Skills (23)

algorithm10

algorithms10

c-language10

zeromq10

distributed-training10

machine-learning10

distributed-systems10

cntk10

data-preprocessing10

deep-learning10

deep-neural-networks10

neural-network10

params10

topic-modeling10

para10

Programming languages (3)

C++CPython

Github contributions (5)

microsoft/Multiverso

Sep 2015 - Jun 2017

Parameter server framework for distributed machine learning

Role in this project:

Back-end Developer

Contributions:1 release, 239 commits, 38 PRs in 1 year 9 months

Contributions summary:Fei's commit introduces initial code and structure for a server component within the "multiverso" project, designed for distributed machine learning. The code includes the implementation of fundamental server functionalities, such as managing rows, processing messages (register, create table, set row, clock, end train), and handling inter-process communication using ZeroMQ. This contribution demonstrates a focus on building core infrastructure components necessary for the parameter server framework.

parameter-servermachine-learningparameterserver-frameworkdistributed-machine-learning

microsoft/LightLDA

Sep 2015 - Dec 2017

Scalable, fast, and lightweight system for large-scale topic modeling

Role in this project:

Back-end Developer

Contributions:42 commits, 5 PRs, 22 pushes in 2 years 3 months

Contributions summary:Fei primarily contributed to the LightLDA project by implementing and refining core back-end functionalities related to data preprocessing and model training. These contributions include the creation of the `dump_binary.cpp` file for converting LibSVM data, and modifications to files like `trainer.cpp`, `data_block.cpp`, and `alias_table.cpp`, which are integral to the LDA model's training process. They also added word count functionality for preprocessing, enhancing the project's utility for topic modeling tasks. These changes improved the efficiency and functionality of the LightLDA system.

scalablepythonldapartitioningmachine-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.

Request Free Trial