Haoyang Fang
Applied Scientist at Amazon Web Services (AWS)
Bellevue, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolHaoyang Fang is an Applied Scientist at AWS AI in Bellevue with four years of industry experience building and refining production-ready machine learning systems. He holds an MS in Computer Vision from Carnegie Mellon and a double first-class honors degree in Pure Math and Computer Science from HKUST, blending strong theoretical chops with practical ML engineering. At AWS he focuses on model distillation and multilingual NLP workflows, contributing notable examples and improvements to the widely used AutoGluon project (including GLUE and PAWS-X distillation pipelines). His background includes internships at Amazon and Samsung R&D, reflecting hands-on experience across research and product teams. Colleagues describe him as a meticulous implementer who surfaces subtle model-loss and metric handling improvements that simplify real-world ML adoption.
3 years of coding experience
Master of Science in Computer Vision, 4.17/4.3, Master of Science in Computer Vision, 4.17/4.3 at Carnegie Mellon University School of Computer ScienceHong Kong University of Science and Technology (HKUST)
Github Skills (16)
pytorch10
machine-learning10
deeplearning-ai10
deep-learning10
automated-machine-learning10
python10
automl10
ensemble-learning9
transfer-learning9
forecast9
forecasting9
nlp8
time-series8
natural-language-processing8
computer-vision7
Programming languages (1)
Github contributions (5)
autogluon/autogluon
Jun 2022 - Jan 2023
Fast and Accurate ML in 3 Lines of Code
Role in this project:
ML Engineer & Data Scientist
ML Engineer & Data ScientistContributions:425 reviews, 34 commits, 181 PRs in 7 months
Contributions summary:Haoyang primarily contributed to the development of an example for automm distillation within the autogluon/autogluon repository. Their work involved implementing a distillation example on GLUE datasets, which included adding a README file and updating the code for a new version. They also made various code improvements, such as automatically inferring the soft label loss function, allowing metrics input to be a string, and adding distillation loss features. The user also focused on the multilingual distillation example on PAWS-X dataset and was involved in refactoring and refactoring the models.
forecastingimage-textmlppythonmeta-learning
FANGAreNotGnu/autogluon
Jun 2022 - Jan 2023
AutoGluon: AutoML for Image, Text, and Tabular Data
Contributions:437 commits, 767 pushes, 74 branches in 6 months
pytorchimage-textpythondata-sciencedeep-learning
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.