Albert Xiao
New York City Metropolitan Area United States
Summary
🤩
Rockstar🎓
Top SchoolAlbert Xu is a quantitative trader and former NLP researcher who blends production C++ trading systems with deep learning expertise. He transitioned from leading instruction for UC Berkeley's massive CS 61A course—overseeing thousands of students and hundreds of TAs—to research on ML uncertainty and NLP under Prof. Dan Klein and at USC. On GitHub he has contributed substantive modules to the fastNLP framework (1-D convs, K-max pooling, trainer updates), reflecting hands-on experience building extensible NLP tooling. At Akuna Capital he focuses on pricing research and low-latency implementation, applying both statistical modeling and software engineering rigor. He’s comfortable moving between academia and finance, with a track record of shipping research-driven code into reusable frameworks. Based in Los Angeles, he combines teaching, open-source contributions, and quant trading to tackle real-world ML and systems challenges.
11 years of coding experience
3 years of employment as a software developer
West Windsor-Plainsboro High School SouthBachelor of Science - BS Computer Science, Bachelor of Science - BS Computer Science at University of Illinois Urbana-ChampaignMaster of Science in Computer Vision, Master of Science in Computer Vision at Carnegie Mellon UniversityDual Enrollment Mathematics, Dual Enrollment Mathematics at Princeton University
Github Skills (10)
mask-rcnn10
faster-rcnn10
pytorch10
deep-learning10
python10
natural-language-processing10
nlpjs9
nlp9
text-classification8
text-processing7
Programming languages (4)
Github contributions (5)
FudanNLP/fastNLP
Jul 2018 - Sep 2018
fastNLP: A Modularized and Extensible NLP Framework. Currently still in incubation.
Role in this project:
ML Engineer
ML EngineerContributions:11 commits, 6 PRs, 1 push in 2 months
Contributions summary:Albert contributed significantly to the `fastnlp` framework by implementing and modifying modules related to 1-D convolutions, max-pooling, and average pooling. They added functionalities like K-max pooling and integrated these components into the existing architecture. Furthermore, they updated the trainer for classification tasks by introducing new modules and adjustments to the training process.
natural-language-processingdeep-learningnlp-librarynlp-parsingchinese-nlp
keezen/keezen.github.io
Nov 2015 - Aug 2022
Contributions:32 pushes, 1 branch in 6 years 10 months
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.