Dongyang Fan
Doctoral Researcher at EPFL
Lausanne Metropolitan Area Switzerland
Summary
🤩
Rockstar🎓
Top SchoolDongyang Fan is a doctoral researcher at EPFL focused on data-efficient large language models and responsible AI, with a strong publication record including NeurIPS, ICML, and ICLR. He combines large-scale LLM pretraining, modular ML design, and rigorous evaluation, and has driven measurable gains in applied settings (e.g., a 1.6× math-metric boost from better synthetic data at DatologyAI). Prior roles span deep learning for robotics and transformer/VAE and GNN work on industrial sensor data where he uncovered leaks far below prior detection thresholds. Trained in statistics at ETH Zürich and with international exposure in Spain and China, he blends theoretical rigor with practical systems engineering. Colleagues describe him as rooted but adaptable, reflected in a career that moves fluid research into high-impact applied outcomes.
3 years of coding experience
1 year of employment as a software developer
Bachelor's degree, Traffic Engineering, Bachelor's degree, Traffic Engineering at Tongji UniversityDoctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at EPFLMaster's degree, Statistics, Master's degree, Statistics at ETH ZürichExchange semester, Traffic Safety Engineering, Exchange semester, Traffic Safety Engineering at Universidad Politécnica de Madrid
Chinese, English, French, German
Github Skills (3)
optimization8
machine-learning8
data-science6
Programming languages (2)
Github contributions (5)
fan1dy/collaboration-consensus
May 2023 - Sep 2024
Contributions:4 pushes, 1 branch in 1 year 3 months
fan1dy/PFL-Non-IID
Nov 2022 - Apr 2023
Personalized federated learning simulation platform with non-IID and unbalanced dataset
Contributions:7 pushes in 5 months
personalizeddatasetmachine-learningsimulationpersonalized-federated-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.