Winston Quock
Principal Machine Learning Engineer at Zillow
San Ramon, California, United States
Summary
👤
Senior🎓
Top SchoolWinston Quock is a Principal Machine Learning Engineer based in San Ramon, CA with nine years focused on server-side systems and extensive experience building large-scale, highly available web, API, and big-data processing platforms. He has led platform and architecture efforts at Zillow, Kareo, Blurb, and Smule, improving reliability and developer productivity across microservices, media pipelines, and transactional workflow systems. Winston combines deep backend engineering—Java/JEE, distributed systems, and scalable data processing—with practical ML cluster expertise, contributing to the widely used TensorFlowOnSpark project to improve distributed TensorFlow on Spark. Known for turning flaky production services into zero-downtime components, he emphasizes operational resilience and performance engineering. He mentors engineering teams and drives architectural consolidation that reduces complexity while enabling scale. A UC Berkeley computer science graduate, Winston blends pragmatic engineering with a track record of shipping robust, production-grade systems at internet scale.
9 years of coding experience
22 years of employment as a software developer
Bachelor of Arts (B.A.) Computer Science, Bachelor of Arts (B.A.) Computer Science at University of California, Berkeley
Chinese
Github Skills (9)
apache-spark10
machine-learning10
cluster-computing10
tensorflow10
python10
clustering10
spark8
mlops8
scala8
Programming languages (5)
Github contributions (5)
yahoo/TensorFlowOnSpark
Dec 2017 - Jan 2018
TensorFlowOnSpark brings TensorFlow programs to Apache Spark clusters.
Role in this project:
ML Engineer
ML EngineerContributions:7 commits, 3 PRs, 5 comments in 1 month
Contributions summary:Winston contributed to the TensorFlowOnSpark project by implementing and refining features related to running TensorFlow models on Apache Spark clusters. Their work included enabling the execution of parameter server (PS) nodes on the Spark driver for efficient resource utilization, as well as daemonizing PS threads and processes. The changes spanned multiple files, including core components like TFCluster, TFSparkNode, and example scripts, demonstrating a focus on cluster management and distributed TensorFlow training workflows. The user also made additions to command-line arguments to the example scripts for increased flexibility and control.
pythonclustersmachine-learningapachespark
zillow/taogpt
Apr 2024 - Sep 2024
TaoGPT: From deep learning to deep thinking - A hybrid System 1 and System 2 AI Agent running on top of LLM
Contributions:10 PRs, 8 pushes, 6 branches in 5 months
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.