Zongheng Yang
Something New at Something new
Berkeley, California, United States
Summary
🤩
Rockstar🎓
Top SchoolZongheng Yang is a software engineer and researcher with 13 years of experience building scalable distributed systems and ML infrastructure, currently leading development on SkyPilot, an open-source platform for running AI workloads across Kubernetes and 15+ clouds. He holds a PhD from UC Berkeley where his dissertation on ML for query optimization earned a SIGMOD Jim Gray Doctoral Dissertation Award honorable mention and produced systems like Balsa, NeuroCard, and Naru. Zongheng has deep backend expertise demonstrated by contributions to Apache Spark (SQL engine and join optimizations), Ray, and TensorFlow at Google Brain, where he improved checkpointing efficiency and worked on notable speech synthesis models. He focuses on cost- and runtime-aware optimization algorithms for cloud workloads, often improving core DAG and resource-allocation logic rather than surface tooling. Based in Berkeley, he blends rigorous research with production-grade engineering and maintains an active open-source presence at skypilot-org and on GitHub.
13 years of coding experience
11 years of employment as a software developer
Doctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at University of California, Berkeley
English, Chinese, German, Chinese
Github Skills (38)
algorithm10
code-optimization10
algorithms10
apache-spark10
spark10
cloud-computing10
redis10
python10
r10
testing10
big-data10
machine-learning10
data-serialization10
distributed-systems10
networkx10
Programming languages (14)
Github contributions (5)
skypilot-org/skypilot
Aug 2021 - Jan 2023
SkyPilot: Run AI and batch jobs on any infra (Kubernetes or 15+ clouds). Get unified execution, cost savings, and high GPU availability via a simple interface.
Role in this project:
Back-end Engineer
Back-end EngineerContributions:4 releases, 1940 reviews, 207 commits in 1 year 5 months
Contributions summary:Zongheng made significant contributions to the core optimization logic within the SkyPilot project, focusing on optimizing for cost and runtime. Their work included refactoring and enhancing existing functions like _egress_cost, as well as introducing new features such as the addition of dummy source and sink nodes for more efficient DAG manipulation. They also made improvements related to resource allocation and handling of different hardware configurations to align with the project's objectives of cloud-based workload optimization. The user demonstrated a deep understanding of the project's core algorithms and concepts, enhancing the tool's performance and efficiency.
pythonspot-instancesserverlesstpuml-infrastructure
amplab-extras/SparkR-pkg
Sep 2013 - Apr 2015
R frontend for Spark
Role in this project:Back-end Developer
Back-end DeveloperContributions:131 commits, 27 PRs, 28 pushes in 1 year 7 months
Contributions summary:Zongheng's primary focus was on developing the R front-end for Spark, implementing core functionalities. They initialized a JavaSparkContext and implemented the textFile() and parallelize() functions. Furthermore, the user added and refined essential features such as collect() and take(), and integrated S4 classes to manage RRDD, demonstrating strong expertise in extending the SparkR library.
apache-sparkfrontendrspark
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.