German Shegalov
Principal Distributed Systems Engineer at NVIDIA
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolGerman Shegalov is a Principal Distributed Systems Engineer with 13 years of experience building high-performance data and ML infrastructure, currently accelerating Apache Spark SQL on GPUs at NVIDIA. He has deep expertise in distributed systems, databases, and ML pipelines from senior engineering roles at Salesforce (TransmogrifAI), Twitter, Oracle, and MapR. A prolific open-source contributor, his work spans core projects like Apache Spark, Hadoop, Spark RAPIDS, and cuDF, where he focused on stability, cross-platform compatibility, GPU acceleration, and subtle correctness fixes (deadlock resolution, multiclass classification handling, and serialization/metrics improvements). German combines research rigor—doctoral-level training and early academic work on exactly-once semantics—with pragmatic production engineering, often touching build systems, shims, and test automation. Based in San Francisco, he quietly bridges low-level performance tuning and large-scale ML workflow reliability, with a knack for surfacing non-obvious issues that improve long-term maintainability.
13 years of coding experience
20 years of employment as a software developer
Diplom-Informatiker, Computer Science, 1.3, Diplom-Informatiker, Computer Science, 1.3 at Universität des SaarlandesDoktor der Ingenieurwissenschaften, Computer Science, Doktor der Ingenieurwissenschaften, Computer Science at Saarland University
Russian, English, German
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (56)
unit-testing10
hadoop-mapreduce10
apache-spark10
spark10
c-language10
cascading10
pytest10
back-end-development10
apache-hadoop10
testing10
dataframes10
memory-management10
big-data10
cudf10
machine-learning10
Programming languages (10)
Github contributions (5)
NVIDIA/spark-rapids
Jan 2021 - Jan 2023
Spark RAPIDS plugin - accelerate Apache Spark with GPUs
Role in this project:
Back-end Developer & Test Automation Engineer
Back-end Developer & Test Automation EngineerContributions:1863 reviews, 168 commits, 454 PRs in 2 years
Contributions summary:German contributed to the development and testing of the Spark RAPIDS plugin, focused on accelerating Apache Spark with GPUs. Their work involved adding and modifying code related to the shims, which are necessary for compatibility with different Spark versions. Furthermore, the user added and modified unit tests to verify the functionality of the plugin, with a focus on ensuring the correct behavior of various features, including division by zero and the performance of the code.
gpu-accelerationgpuapachebig-dataspark
salesforce/TransmogrifAI
Aug 2018 - Nov 2020
TransmogrifAI (pronounced trăns-mŏgˈrə-fī) is an AutoML library for building modular, reusable, strongly typed machine learning workflows on Apache Spark with minimal hand-tuning
Role in this project:Back-end Developer & DevOps Engineer
Back-end Developer & DevOps EngineerContributions:1 release, 27 reviews, 21 commits in 2 years 3 months
Contributions summary:German primarily contributed to the core codebase of TransmogrifAI, focusing on improving the stability and maintainability of the system. Their contributions involved fixing potential deadlocks, creating OS-neutral file system paths, and refactoring code related to data handling within the core workflows. In addition, the user addressed multiclass classification issues, enhancing the accuracy and robustness of the machine learning model selection process. Furthermore, the user's involvement extended to improving metrics aggregation, which included runtime spark metrics and model serialization formats.
pythonsalesforcetransformationstransmogrifyestimators
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.