"Machine learning or deep learning? None, I just keep learning"
-- Luke Qi
-- Luke Qi
Hi, this is Luke Qi!
I am currently finishing my Master’s of Science in Data Science(MSDS) at University of San Francisco, where I have developed a strong programming and data warehouse skills and become passionate about applying machine learning methods to solve business problems.
Prior to that, I graduated with a BS in Statistics and a BA in Mathematics at University of Minnesota - Twin Cities, where I built a solid knowledge of statistics and mathematics, and realized the how powerful data could be in decision making.
As a Data Scientist Intern at Ubisoft, I am responsible for helping the marketing team boost revenue and optimize pricing strategy by using predictive machine learning model to analyze users' behavior.
The followings are some of my projects that I completed so far that I have been interested in.
Click "more" for details and source code on github.
A python package that generates detailed machine learning model evaluation metrics which are useful in industry
(Author, Maintainer, 178 stars and 31 forks)
A end-to-end machine learning model pipeline in game industry to predict monetizers.
(Python, SQL, Hive, Pipeline, Giraffez)
Predict Santander customer transcation in the future with stacked boosting models.
Rank 44/9038 on Kaggle Competition.
(Python, Pipeline, Augmentation, Stacking)
Time Series forecast of Canadian bankruptcy rate with macroeconomic indicator.
(R, Holt-Winters, SARIMA, VARX)
A neural network built from scratch to predict the bike share at WA D.C.
(Python, DNN, Hyperparameter Tuning)
White wine quality classification using ensemble machine learning models.
(Python, Resampling, Imbalanced Data)
Leverage ML algorithms through ETL Pipeline to predict Single-Vehicle Traffic Accident Severity.
(AWS, Spark, MongoDB, s3)A web app which interacts with a RNN model performing sentiment analysis on movie reviews.
(SageMaker, RNN, Model Deployment)Leverage ML models through pipeline to process large-scale data to classify purchased user.
(Python, GCP, AWS, Big Data)A web app to classify real-world, user-supplied dog/cat/human images.
(Pytorch, ResNet50, VGG16, CNN)A LSTM model to generate my own Seinfeld TV scripts.
(Pytorch, RNN, LSTM, Embedding)A DCGAN model to generate new images of faces that look as realistic as possible.
(Pytorch, GAN, Convolutional Layer)