Nice to meet you,
I'm James Fu 👋Hello! I'm James Fu 👋
I'm James Fu 👋
Data & Supply Chain Analyst
James Fu
MSDS @ UT Austin
|
I recently completed my MS in Data Science from UT Austin. Prior to that, I earned my BS as a proud Bruin studying Computational Systems and Biology @ UCLA.
My background bridges advanced data analytics and supply chain operations, currently working on improving inventory planning + automating long workflows. Whether I'm coding for fun or professionally, I thrive on solving complex challenges through analytics and ML.
A fun fact is that previously taught computer science and math. Check out my the page below to see some of my favorite shots from my Ricoh!
Education
M.S. Data Science
2024–2026University of Texas at Austin
Focus on data analysis, predictive modeling, and statistical methods.
Explore →B.S. Computational Systems and Biology
2020–2024University of California, Los Angeles
Concentration in Data Science. Focus on health data analytics and ML, working on spatial transcriptomics at the Zipursky Lab (23-24), PM on the Biomedical Engineering Society's Build Team (23-24), front end development @ DevX (22-23).
Explore →Projects
Gene Expression and Subtype Analysis of Astrocytes in the Mouse Brain
Developed a pipeline using unsupervised machine learning to identify six astrocyte subtypes through spatial clustering of spatial transcriptomic data from the Allen Mouse Brain Atlas, discovered 104 astrocyte-specific genes through differential expression analysis with Bonferroni correction across 10+ million cells.
Enhancing Robustness in Natural Language Inference Models
Followed-up on methods explored in Swayamdipta et al. (2020) to improve NLI robustness. Fine-tuned ELECTRA-small using dataset cartography and contrast sets with distractors, reweighting hard-to-learn examples during training to reduce artifacts. Achieved an 8% accuracy increase on novel contrast sets.
Developing Faircare for Transparent Pricing
Built a healthcare cost modeling website leveraging synthetic data augmentation and data analysis to make accurate cost predictions. Increasing volume of MEPS dataset by 253%, this tool provides intuitive visualizations summarizing fair treatment cost estimates, ensuring transparency for all users, regardless of insurance coverage.
Dementia Care via Voice Assistance and Facial Recognition
Won 1st place by building Dotmentia, an assistive AI system that leverages facial recognition (OpenCV) and Google Assistant integration (DialogFlow) to help dementia patients recognize family members and those around them in real-time, enhancing their daily interactions and independence.
Skills
Python
Pandas
SQL
SAP ERP
Power Automate
Tableau
PyTorch
Git
React
Contact Me
Get in touch
Always open to connect or learn about new opportunities.
jamesfup@gmail.com
San Jose, CA