Sewon Min
Assistant Professor at UC Berkeley Electrical Engineering & Computer Sciences (EECS)
Berkeley, California, United States
Summary
🤩
Rockstar🎓
Top SchoolSewon Min is an assistant professor and researcher in Berkeley with roughly a decade of experience building and optimizing NLP and retrieval systems. Currently at UC Berkeley EECS and a research scientist at AI2, he moved from a PhD at the University of Washington into roles that blend core research with production-focused engineering. His open-source contributions include backend improvements to Facebook Research’s Dense Passage Retriever—work that streamlined indexing, gzipped data handling, and robust file management for large-scale open-domain QA. Known for pragmatic code that bridges research prototypes and deployable infrastructure, he’s comfortable across modeling, data pipelines, and systems-level tooling. He prefers email contact over LinkedIn and maintains a personal site at sewonmin.com.
10 years of coding experience
PhD student Computer Science & Engineering, PhD student Computer Science & Engineering at University of WashingtonBachelor of Computer Science and Engineering Computer Science and Engineering, Bachelor of Computer Science and Engineering Computer Science and Engineering at Seoul National UniversityHigh School Diploma, High School Diploma at Gyeonggi Science High School for the Gifted (경기과학고등학교)
Github Skills (5)
file-handling10
indexing10
python10
data-retrieval10
gzip9
Programming languages (5)
Github contributions (5)
facebookresearch/DPR
Jun 2020 - Jun 2020
Dense Passage Retriever - is a set of tools and models for open domain Q&A task.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:18 commits, 7 PRs, 15 pushes in 5 days
Contributions summary:Sewon primarily contributed to the project by modifying core Python scripts related to data retrieval and indexing, specifically within the context of a dense passage retriever for open-domain question answering. Their commits involved updates to the `dense_retriever.py` file, focusing on index management (saving/loading), and file handling. Additionally, the user implemented options for handling gzipped files in data loading and downloading utilities and made several changes and optimizations to existing code.
open-domainretrieverdomainpassagedense
shmsw25/DecompRC
Jun 2019 - Jul 2020
An original implementation of ACL 2019, "Multi-hop Reading Comprehension through Question Decomposition and Rescoring"
Contributions:13 commits, 5 pushes, 2 branches in 1 year 1 month
pytorchnlpdecompositionquestionhop
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.