John Zedlewski
VP, Accelerated Data Science
Berkeley, California, United States
Summary
🤩
Rockstar🎓
Top SchoolJohn Zedlewski is a VP-level engineering leader with 14+ years driving GPU-accelerated data science and AI products at NVIDIA, where he leads the RAPIDS open-source effort that brings Pandas-, Spark-, and SQL-friendly APIs to CUDA performance. He has progressed from deep learning research for autonomous driving to architecting end-to-end ML and retrieval systems—shipping work spanning UMAP improvements in cuML to performance benchmarking notebooks that helped demonstrate cuML vs scikit-learn. Based in Berkeley, he combines academic rigor from Harvard and Princeton with hands-on kernel and optimization experience, enabling teams to translate novel algorithms into production-grade, GPU-first tooling. Known for bridging research and developer ecosystems, he focuses on practical performance wins—like accelerating GNNs, RAG retrievers, and agentic frameworks—that let users build AI applications faster.
14 years of coding experience
12 years of employment as a software developer
MA ABD Economics, MA ABD Economics at Harvard UniversityAB Computer Science, AB Computer Science at Princeton University
Github Skills (16)
algorithm10
scikit-learn10
machine-learning10
rapids10
machine-learning-algorithms10
cuml10
python10
omap10
imap10
scikit10
gpu-acceleration9
benchmark9
benchmarking9
kmeans9
data-analysis8
Programming languages (7)
Github contributions (5)
rapidsai/cuml
May 2019 - Apr 2021
cuML - RAPIDS Machine Learning Library
Role in this project:
Back-end Developer
Back-end DeveloperContributions:457 reviews, 408 commits, 320 PRs in 1 year 11 months
Contributions summary:John made several commits focused on enhancing the UMAP functionality within the cuML library. Their work included adding notes about unsupported UMAP features, updating documentation to reflect review feedback, and merging branches to integrate new code, showing their contributions to feature improvement and code integration. The user's efforts focused on the `cuml.manifold.umap.pyx` file, which is involved in dimension reduction and machine learning algorithm development.
cudacumlnvidiadata-sciencegpu
rapidsai-community/notebooks-contrib
May 2019 - Nov 2019
RAPIDS Community Notebooks
Role in this project:Data Scientist
Data ScientistContributions:9 commits, 4 PRs, 12 comments in 5 months
Contributions summary:John contributed to the "cuml_benchmarks.ipynb" notebook, focusing on performance benchmarking of cuML algorithms within the RAPIDS ecosystem. Their work involved adding and cleaning benchmark functions for machine learning algorithms such as KMeans, demonstrating their use of cuML for GPU-accelerated machine learning. The user's changes indicate an effort to analyze and compare the performance of cuML algorithms with their scikit-learn counterparts.
jupyter-notebooknotebooksrapids
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.