Dan Xie
Product Manager at Pinterest
Berkeley, California, United States
Summary
🤩
Rockstar🎓
Top SchoolDan Xie is a product manager with a decade of experience building consumer and advertising products, currently at Pinterest after leading TikTok Ads Manager and multiple ByteDance initiatives. He combines a strong data science foundation from UC Berkeley with hands-on machine learning engineering contributions—such as augmenting the widely used robustness repo for corrupted-image benchmarks—to bridge research and product execution. At TikTok he drove major lifts in ad completion, adoption, and revenue, launched shopping ads and mobile experiences, and led multi-market rollouts that scaled advertiser growth and monetization. His background spans from development practice and applied data science to international policy and sustainability work, giving him a systems-level view of product-market fit. Notably, he has experience shipping both technical dataset tooling and high-impact product features that translate ML robustness insights into production-ready ad experiences.
10 years of coding experience
4 years of employment as a software developer
Master's degree Master of Development Practice, Master's degree Master of Development Practice at University of California, BerkeleyBachelor's degree International Relations and Affairs, Bachelor's degree International Relations and Affairs at University of International Business and Economics
Github Skills (7)
computer-vision10
pytorch10
machine-learning10
deeplearning-ai10
deep-learning10
numpy8
python8
Programming languages (7)
Github contributions (5)
hendrycks/robustness
Apr 2018 - Aug 2022
Corruption and Perturbation Robustness (ICLR 2019)
Role in this project:
ML Engineer
ML EngineerContributions:64 commits, 9 PRs, 59 pushes in 4 years 5 months
Contributions summary:Dan contributed to the `robustness` repository, which focuses on corruption and perturbation robustness in the context of deep learning and computer vision. Their commits include the addition and modification of data loading scripts, dataset classes, and image corruption methods, indicating a focus on data preparation and augmentation techniques. The user also introduced new files related to creating corrupted versions of CIFAR and ImageNet datasets, directly contributing to the core functionality of the project. The overall changes suggest that the user was involved in generating and handling corrupted image data to evaluate model robustness.
imagenetml-safetypytorchiclrcorruption
hendrycks/natural-adv-examples
Jun 2019 - Feb 2021
A Harder ImageNet Test Set (CVPR 2021)
Contributions:16 commits, 15 pushes, 8 comments in 1 year 9 months
adversarial-attacksadversarial-exampledomain-generalizationimagenetml-safety
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.