Yuan Hui, Ph.D. Candidate
Yuan Hui
Data Scientist
Tel: 716-***-****
Email: *******@*******.***
**** ******** **. ***.*, Tonawanda, NY, 14150
LinkedIn Github Personal Webpage
An interdisciplinary engineer and researcher with 5 years’ experience in data science, machine learning and computational fluid dynamics. Special interests are to build and optimize predictive models and to discover the association relationships in data.
Skills
Language: Python, R, SQL, MATLAB
Libraries: Numpy, Scipy, Pandas, Pytorch, Keras, Scikit-learn, OpenCV in Python; glmnet, xgboost, caret, rpart, neuralnet, gbm in R
Others: Git/GitHub, AWS Sagemaker, GCP BigQuery
Projects
LSTM: Predict Watershed Nutrients Loading (publication in prep.) – details here May 2019 – Present
Train and optimize LSTM model with feature selection and hyperparameter tuning
MSE is 0.13, which is superior than ARIMAX model 0.25
Libraries: Scipy, Pandas, Pytorch
Unsupervised : Time Series Anomaly Detection (publication in prep.) – details here Mar 2019 – Present
Build PCA, hierarchical and K-Means clustering on long term nutrients time series
Find that high nutrient in late spring with high temporal variance increases eutrophication possibility
Libraries: Tslearn, Sklearn, Scipy, Pandas, Numpy ARIMAX: Market Share Analysis and Forecasting Model on Weekly Sales Oct 2019 – Feb 2020
Create ARIMAX models for weekly sales time series with 7 lagged exogenous variables
RMSE for the training error is 166.8 sales. Average test error rate is 6.1%.
Libraries: Pandas, Numpy, Statsmodels
CNN: Image Classification – details here Jun 2019
Apply cascade classifiers to detect human face with accuracy of 98%; Apply pre-trained VGG-16 to recognize dogs with accuracy of 96%
Use transfer learning of pre-trained Resnet50 on dog classification to increase precision of base model
(three convolutional layers) from 11% to 81%
Libraries: Numpy, OpenCV, Pytorch
Regression and Classification: House Prices Prediction– details here Nov 2018 – Dec 2018
Exploratory Data Analysis to fill missing values for 7 features and their correlation analysis
Regression (Lasso, Ridge, XGBoost) for price prediction with MAPE less than 10%; Classification
(Random Forest and SVM) to classify high/low prices with accuracy of 60%
Libraries: glmnet, xgboost, caret, rpart, neuralnet, gbm Experience
Data Scientist Intern, ACV Auctions, Buffalo, NY Sep 2019 – Feb 2020
Detect car engine audios for tick/knock with recall of 0.65 using XGBoost; LSTM improve recall to 0.89
Object detection on car images with accuracy of 0.99; Google OCR text retrieval with accuracy of 0.96.
Great team collaboration and documentation with Git Deep Learning Nanodegree Mentor, Udacity Dec 2019 – Present
Mentor more than 50 students on deep learning computer vision projects
The mentored projects include image classification using CNN, RNN applications on NLP, image to image translation using GANs
2
Yuan Hui, Ph.D. Candidate
Graduate Student Chapter Chair, Environmental and Water Resource Institute Mar 2018 – Sep 2019
Found the student chapter and achieved $1500 funding
Lead 7 graduate student activities in water resource field Education
Ph.D. Civil Engineering, University at Buffalo, NY Expected in Sep 2020 M.S. Data Science, University at Buffalo, NY Feb 2020 M.S. Hydro-informatics, University of Nice Sophia Antipolis, France Sep 2014 B.E. Hydraulic and Hydro-power Engineering, Chongqing Jiaotong University, China Jun 2012 Selected Publications
Hui et al. Time series analysis using unsupervised learning on tributary phosphorus loading and their effects on nearshore water eutrophication in Lake Ontario. (In prep.) Hui et al. Mass balance analysis and calculation of wind effects on heat fluxes and water temperature in a large lake, Journal of Great Lakes Research. (2018) 44 (6), 1293-1305. https://doi.org/10.1016/j.jglr.2018.09.003 Selected Honors & Awards
Graduate Leadership Award, University at Buffalo, NY 2019 First Place of Graduate Technical Paper, Environmental & Water Resource Institute 2019 Dean’s Scholarship, University at Buffalo, NY 2015 Erasmus Mundus Master Degree Scholarship, European Union 2012-2014 Chinese National Scholarship, Department of Education in China 2010