Pei Sun
Software Engineer at Google Deepmind
Summary
🤩
RockstarPei Sun is a software engineer with 10 years of experience focused on back-end systems and machine learning data tooling. She has contributed meaningful fixes and performance improvements to Alluxio—optimizing UFS block reads, packet IO, and request cancellation—to make cloud data orchestration more stable and efficient. On the Waymo Open Dataset she addressed bounding-box handling, upgraded examples for TF2 and Python 3.8 compatibility, and refined build tooling, reflecting a practical ML engineering skillset. Comfortable across systems-level Java networking and Python ML pipelines, she blends reliability-driven engineering with pragmatic dataset and tooling improvements. An attention to testing and end-to-end correctness shows a bias for shipping robust, verifiable changes in high-impact open-source projects.
10 years of coding experience
Github Skills (10)
javas10
tensorflow10
netty10
python10
autonomous-driving10
java10
datasets10
computer-vision9
dockers7
docker7
Programming languages (4)
Github contributions (5)
waymo-research/waymo-open-dataset
Dec 2019 - Mar 2021
Waymo Open Dataset
Role in this project:
ML Engineer
ML EngineerContributions:5 reviews, 14 commits, 13 PRs in 1 year 2 months
Contributions summary:Pei primarily contributed to the Waymo Open Dataset, a dataset for autonomous driving. Their work involved fixing issues related to the handling of bounding boxes, ensuring the correct data processing. The user also made updates to example prediction files and refactored code to be compatible with TF2. Furthermore, there are commits showing the addition of Python 3.8 support and changes to the build and Dockerfile configuration.
datasetopen-datasetautonomous-drivingwaymo-open-dataset
Alluxio/alluxio
Apr 2016 - Jun 2017
Alluxio, data orchestration for analytics and machine learning in the cloud
Role in this project:Back-end Developer
Back-end DeveloperContributions:1094 commits, 350 PRs, 73 pushes in 1 year 2 months
Contributions summary:Pei's primary contributions involve implementing and modifying the `BlockInStream` and related classes to optimize the handling of data packets, stream operations and overall UFS block reading, including handling UFS read requests and making sure the client can effectively cancel read requests. The user also made changes to `NettyPacketReader` and `NettyPacketWriter` to support the new workflow. The user implemented tests to verify their changes, ensuring the file system correctly reads and writes data from UFS. These changes enhance performance and stability.
data-analysisvirtual-distributed-filesystemanalyticsorchestrationdata-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.