Haofei Xu
Data Engineer
Missouri, United States
Summary
🤩
Rockstar🎓
Top SchoolHaofei Xu is a data engineer with a decade of cross-disciplinary experience combining UX sensibilities, large-scale data pipelines, and applied machine learning. Currently building auto-scalable ETL and data lake solutions at J&K Scientific and running municipal data pipelines at the University of Michigan, he turns messy web and video sources into queryable lakes using AWS, Spark, Trino, and Apache Iceberg. His background in UX and psychology informs a user-centered approach to data products and a knack for surfacing the right information at the right time. Contributions to the unimatch repository demonstrate hands-on ML engineering on state-of-the-art 3D vision/depth estimation code, including improving inference and output formats for reproducible evaluation. Based in Missouri, he blends research rigor from ETH/University of Tübingen collaborations with practical production engineering to deliver scalable, interpretable analytics.
10 years of coding experience
1 year of employment as a software developer
Master of Science in Information Information Science, Master of Science in Information Information Science at University of MichiganBachelor's degree Social Psychology Minor in Computer Science and Applied Statistics, Bachelor's degree Social Psychology Minor in Computer Science and Applied Statistics at University of Illinois Urbana-ChampaignHigh School Diploma, High School Diploma at Yangzhou High School of Jiangsu Province
English, Chinese
Github Skills (7)
depth-estimation10
computer-vision10
pytorch10
evaluation9
data-loading9
transformer8
python7
Programming languages (4)
Github contributions (5)
autonomousvision/unimatch
Nov 2022 - Feb 2023
[TPAMI'23] Unifying Flow, Stereo and Depth Estimation
Role in this project:
ML Engineer
ML EngineerContributions:27 commits, 4 PRs, 30 pushes in 3 months
Contributions summary:Haofei's contributions primarily focused on updating and refining the depth estimation components of the unimatch project. This included modifications to data loading and preparation scripts for demonstration datasets, as well as enhancements to the evaluation and inference pipelines. Furthermore, the user implemented the ability to save disparity predictions as .pfm files, suggesting improvements to the model's output and evaluation processes. The commits indicate a hands-on role in refining the model's functionality and output formats.
stereomatchingoptical-flowdepthestimation
haofeixu/aanet
Mar 2020 - Nov 2022
AANet: Adaptive Aggregation Network for Efficient Stereo Matching, CVPR 2020
Contributions:20 commits, 3 PRs, 19 pushes in 2 years 8 months
aggregationstereoadaptivecvpr-2020matching
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.