Kevin Zhang
Staff Machine Learning Engineer at Lightspark
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolKevin Zhang is a Staff Machine Learning Engineer based in the San Francisco Bay Area with 11 years of experience building production ML systems and low-latency AI services. He combines deep academic training from MIT (BS & M.Eng in EECS/AI) with hands-on industry impact across Facebook’s Ads Core ML and Lightspark, driving multi-task models, knowledge distillation, semi-supervised learning, and strategic low-latency AI projects. His research background spans synthetic data generation—working with copulas, GANs, and VAEs—and he has contributed practical improvements to open-source copulas tooling, including KS-based goodness-of-fit and better support for truncated/beta distributions. Kevin has repeatedly bridged research and product: from decentralized privacy-preserving ML marketplaces at CSAIL to high-throughput emotion-detection and real-time ad-targeting systems. He’s comfortable leading small research teams and shipping at scale, with a track record of translating statistical rigor into robust, deployable models. A subtle but telling thread in his career is an emphasis on data fidelity—synthesizing, modeling, and validating complex distributions to make ML systems both accurate and reliable.
11 years of coding experience
8 years of employment as a software developer
Master of Engineering - MEng Artificial Intelligence, Master of Engineering - MEng Artificial Intelligence at Massachusetts Institute of TechnologyThe Harker School
English
Github Skills (8)
scipy10
probability-distribution10
statistical-models10
python10
distributions10
data-science10
testing9
machine-learning9
Programming languages (8)
Github contributions (5)
sdv-dev/Copulas
Feb 2020 - Jan 2021
A library to model multivariate data using copulas.
Role in this project:
Data Scientist
Data ScientistContributions:4 reviews, 31 commits, 32 PRs in 11 months
Contributions summary:Kevin primarily contributed to the development and testing of the `copulas` library for modeling multivariate data using copulas. Their work included the implementation of the Kolmogorov-Smirnov (KS) statistic for evaluating the goodness-of-fit of copula models. They also addressed and fixed issues within the code, improving the accuracy of distribution fitting, particularly for truncated normal distributions and implementing support for new distribution types such as the beta distribution.
pythontabular-datagenerative-modelmachine-learningdataset-generation
k15z/CsEdu-3-Perceptron
Mar 2015 - Dec 2015
Contributions:23 commits, 13 pushes, 1 branch in 8 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.