Alfred Xu
Software Design Engineer at Microsoft ShenZhen (MACH)
Shenzhen, Guangdong Province, China
Summary
🤩
Rockstar🎓
Top SchoolAlfred Xu is a seasoned Software Design Engineer based in Shenzhen with 11 years of experience building backend systems, test tooling, and engineering automation at Microsoft and Foxconn. He combines a strong C/C++/C# foundation from industrial CAM and CAD tool development with modern contributions to high-profile open-source ML and data projects like XGBoost, cuDF, Kubeflow Katib and NVIDIA Spark-RAPIDS. His open-source work shows a practical focus on JVM interoperability, decimal handling on GPUs, distributed prediction correctness, and extensible plugin architectures—skills that translate to robust production integrations. At Microsoft he designs test station tools end-to-end, bringing hardware-aware software design and test automation expertise. He’s comfortable bridging ML model tooling and large-scale data processing, and often surfaces runtime and compatibility improvements rather than just feature additions. Outside work he hints at a playful creative side—his GitHub bio reads in Japanese as “a foolish me just sings”—suggesting a curious, human approach to engineering.
11 years of coding experience
7 years of employment as a software developer
Bachelor's degree, Mechanical Engineering, Bachelor's degree, Mechanical Engineering at Xiangtan University
Github Skills (40)
unit-testing10
xgboost10
kubernetes10
decimals10
c-language10
spark10
jvm10
serializer10
python10
data-engineering10
bigdecimal10
cudf10
decimal-numbers10
machine-learning10
java10
Programming languages (8)
Github contributions (5)
NVIDIA/spark-rapids
Nov 2020 - Jun 2022
Spark RAPIDS plugin - accelerate Apache Spark with GPUs
Role in this project:
Back-end Developer & Data Engineer
Back-end Developer & Data EngineerContributions:325 reviews, 96 commits, 154 PRs in 1 year 7 months
Contributions summary:Alfred primarily contributed to enhancing the `nvidia/spark-rapids` project by adding support for decimal data types, addressing various issues in handling them within the Spark RAPIDS plugin. Their work included adding decimal support for `GpuProjectExec` and verifying the shuffle of decimal type data. The user also implemented functionality for `GpuFilter`, `GpuCoalesceBatches`, `GpuSubqueryBroadcast`, `GpuCollectList` and `GpuCollectSet` and added tests for them, as well as adding support for reading decimal columns from parquet files. In addition, they also included support for operations like array_repeat, arrays_zip and providing bug fixes on decimal/string types.
gpu-accelerationgpuapachebig-dataspark
kubeflow/katib
Dec 2019 - Sep 2020
Automated Machine Learning on Kubernetes
Role in this project:Backend Developer
Backend DeveloperContributions:7 reviews, 15 commits, 19 PRs in 8 months
Contributions summary:Alfred primarily contributed to the backend functionality of the Katib project, focusing on enhancements to the suggestion service. Their commits involved adding disk settings to the suggestion configuration, modifying the metrics collector to skip irrelevant lines, and updating the ExperimentStatus to record trial names grouped by stages. They also implemented dynamic job provider and suggestion composer registration, showcasing a strong focus on improving the core capabilities of the Katib system.
xgboosthyperparametermachine-learninglightgbmautoml
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.