Michael Zhu
Founder at Flow Labs
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolMichael Zhu is a founder and software engineer with nine years of experience building AI-first products and cloud-native infrastructure in the San Francisco Bay Area. He combines hands-on ML engineering and DevOps expertise—contributing to notable open-source projects like Ray and SkyPilot—to enable distributed training, multi-cloud execution, and GPU-optimized pipelines. At Atlassian he helped ship core components for an AI search product and drove FedRAMP/FIPS security migrations, demonstrating both product and compliance rigor. He founded Flow Labs, shipping consumer apps including an AI-driven debate trainer and a health tracking app, showing a knack for turning research-grade ML into consumer experiences. A UC Berkeley computer science alum who taught data science courses and optimized ML workflows in research, Michael blends strong pedagogy with practical engineering. Less obvious: he pairs cloud-scale automation with a background in coaching and founding small ventures, so he moves between technical depth and user-focused product growth.
9 years of coding experience
7 years of employment as a software developer
Bachelor of Arts - BA, Computer Science, Bachelor of Arts - BA, Computer Science at University of California, BerkeleyIrvington High School
English, Chinese
Github Skills (25)
pytorch10
ppp10
distributed-training10
cloud-computing10
rllib10
python10
machine-learning10
sac10
reinforcement-learning10
deep-learning10
tensorflow10
data-science9
horovod9
gcp9
aws9
Programming languages (2)
Github contributions (5)
skypilot-org/skypilot
Oct 2021 - Jan 2023
SkyPilot: Run AI and batch jobs on any infra (Kubernetes or 16+ clouds). Get unified execution, cost savings, and high GPU availability via a simple interface.
Role in this project:
ML Engineer & DevOps Engineer
ML Engineer & DevOps EngineerContributions:136 reviews, 287 commits, 43 PRs in 1 year 2 months
Contributions summary:Michael contributed to the development and deployment of machine learning models on various cloud infrastructures. Their work includes implementing a Horovod example, demonstrating the use of distributed training, and setting up infrastructure for AI workloads on Kubernetes or multiple cloud providers. The user was also responsible for setting up and managing cloud storage solutions (S3 and GCS) to handle datasets and model artifacts. Furthermore, the user's contributions extend to automation of tasks such as authentication, code formatting, and the overall streamlining of the deployment pipeline.
pythonspot-instancesserverlesstpuml-infrastructure
ray-project/ray
Jan 2019 - May 2021
Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.
Role in this project:ML Engineer
ML EngineerContributions:261 reviews, 33 commits, 47 PRs in 2 years 4 months
Contributions summary:Michael primarily worked on the Ray RLlib library, contributing significantly to the Proximal Policy Optimization (PPO) and related algorithms like APPO and SAC. Their contributions included implementing features such as gradient clipping and importance sampling, fixing performance issues, and improving the documentation and examples. The user also added support for offline RL and the D4RL dataset.
pythonconsistsruntimetensorflowserving
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.