Wei-lin Chiang
Cofounder, CTO at Arena
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolWei-lin Chiang is a CTO and cofounder with 11 years of experience building AI systems, MLOps tooling, and cloud-native infrastructure from research to production. A Berkeley PhD student turned entrepreneur, he led projects at SkyLab and helped create evaluation platforms like Chatbot Arena while driving research internships at Amazon, Google, Microsoft, and Alibaba. On open source he’s contributed significant MLOps and cloud integrations to SkyPilot—adding TPU VM/Pod support, spot-instance fixes, Azure improvements, and IAM automation—to make large-model workloads portable and cost-effective across 15+ clouds. His background in large-scale linear and graph ML systems (liblinear, Cluster-GCN) underpins a pragmatic approach to performance and cross-platform compatibility. Based in San Francisco, he blends deep research rigor with hands-on engineering to measure and advance AI at the frontier.
11 years of coding experience
Doctor of Philosophy - PhD, EECS, Doctor of Philosophy - PhD, EECS at University of California, BerkeleyMaster's degree, Computer Science, Master's degree, Computer Science at National Taiwan University
Github Skills (23)
c-language10
cloud-computing10
python10
machine-learning10
cloud-infrastructure10
tpu10
gcp10
infrastructure10
mlops10
cprogramming-language10
linear-algebra10
kubernetes9
microsoft-azure9
makefile9
kubernetes-pods9
Programming languages (6)
Github contributions (5)
cjlin1/liblinear
May 2015 - Oct 2020
LIBLINEAR -- A Library for Large Linear Classification
Role in this project:
Back-end Developer
Back-end DeveloperContributions:23 commits, 12 comments in 5 years 6 months
Contributions summary:Wei-lin primarily focused on enhancing the build process and core functionalities of the liblinear library. Their contributions include fixing the Python interface on Windows, refactoring the code to improve performance by combining Xv and XTv operations, and adding a new option for not regularizing the bias. They also updated binaries and refactored code to improve performance. These changes indicate a focus on optimizing the library's performance, cross-platform compatibility, and user experience.
liblineardeep-learningmachine-learninglinearlinear-classification
skypilot-org/skypilot
Sep 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:MLOps Engineer & Cloud Engineer
MLOps Engineer & Cloud EngineerContributions:1 release, 354 reviews, 132 commits in 1 year 4 months
Contributions summary:Wei-lin's commits primarily focused on enhancing SkyPilot's capabilities related to running AI and machine learning workloads, particularly on Google Cloud Platform (GCP) and its TPU resources. Their work involved the implementation of TPU support, including the creation of TPU VMs and Pods, and integration with the SkyPilot infrastructure. They also addressed issues with existing spot instance functionality and added improvements for Azure, encompassing image configuration and disk size customization. Additionally, the user implemented the necessary IAM roles for users.
pythonspot-instancesserverlesstpuml-infrastructure
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.