Edward Pizzi
Software Engineer at Facebook AI
San Francisco, California, United States
Summary
👤
Senior🎓
Top SchoolEdward Pizzi is a software engineer in San Francisco with deep expertise in computer vision, distributed systems, and scalable infrastructure, blending nine years on Google's core infrastructure with startup and research experience. At Facebook AI he focuses on fine-grained visual recognition for commerce—powering visual search, duplicate detection, and product clustering—and has contributed notable improvements to the ClassyVision PyTorch framework for image/video classification. He has end-to-end ML experience from data collection and annotation to production deployment and retrieval, and brings practical metric-learning expertise implemented in PyTorch. Earlier roles at GrokStyle and Leanplum show a track record of shipping recommender systems, robust API/SDK designs, and infrastructure automation at scale. Less obvious: he pairs low-level automation instincts from his SRE background (e.g., automated rollout and load-test systems) with ML engineering, making him effective at deploying reliable, production-grade vision systems.
8 years of coding experience
15 years of employment as a software developer
Bachelor of Arts (B.A.), Computer Science, Bachelor of Arts (B.A.), Computer Science at Rice University
Github Skills (8)
computer-vision10
pytorch10
machine-learning10
distributed-training10
python10
model-optimization9
image-classification9
deep-learning9
Programming languages (4)
Github contributions (5)
facebookresearch/ClassyVision
Dec 2019 - Sep 2021
An end-to-end PyTorch framework for image and video classification
Role in this project:
ML Engineer
ML EngineerContributions:13 commits, 13 PRs, 1 comment in 1 year 9 months
Contributions summary:Edward made significant contributions to the `classyvision` project, focusing on improving the training and evaluation processes for image and video classification models. Their work included implementing buffer synchronization options for distributed training, refactoring the classification task for easier extension, and adding support for losses with learned parameters. They also addressed performance issues, such as fixing the cutmix augmentation and optimizing model profiling.
pytorchimagenetresnetend-to-endmeta-learning
facebookresearch/sscd-copy-detection
Mar 2022 - Jul 2022
Open source implementation of "A Self-Supervised Descriptor for Image Copy Detection" (SSCD).
Contributions:4 commits, 3 PRs, 3 pushes in 3 months
pytorchsuperviseddeep-learningdescriptorobject-detection
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.