Changhong Zhang
Research Assistant Institute For Integrating Statistics In Decision Sciences (I²SDS)
New York City Metropolitan Area United States
Summary
🤩
Rockstar🎓
Top SchoolChanghong Zhang is an applied ML researcher and engineer with nine years of experience building production-ready AI systems and a strong academic foundation in statistics and data science. As a Research Assistant at GWU’s I²SDS he implemented a novel layer-wise GPT distillation using Wasserstein distance that improved perplexity and stability, and contributed to multiple peer-reviewed papers on ML-driven optimization. He designs reproducible, cloud-native ML templates and teaching materials—supporting 150+ students—and has productionized RAG and agentic pipelines (OCR, BM25+FAISS retrieval, reranking) into Dockerized AWS services. Changhong blends competition-grade feature engineering (Top 2% in an AmEx Kaggle challenge) with practical tooling for experiment management, A/B sandboxes, and lightweight serving, and he is particularly focused on LLM optimization, decision-making in dynamic environments, and financial time-series modeling.
9 years of coding experience
Master of Science - MS, Business Analytics, Master of Science - MS, Business Analytics at The George Washington UniversityBachelor of Science - BS, Statistics, Bachelor of Science - BS, Statistics at Guangzhou University
Chinese, Chinese, English
Github Skills (13)
aio9
representation-learning9
async7
server-side7
dpdk7
incremental-learning7
reproducible-research7
pytorch7
seastar5
machine-learning5
deep-learning4
python4
overlays2
Programming languages (4)
Github contributions (5)
zhchuu/continual-learning-reproduce
Apr 2020 - Jun 2022
Continual / incremental / lifelong learning methods implemented by PyTorch. Especially the methods based on memory replay.
Contributions:34 commits, 33 pushes, 1 branch in 2 years 2 months
pytorch
zhchuu/zhchuu.github.io
Jun 2020 - Jul 2025
Contributions:68 pushes, 1 branch in 5 years 1 month
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.