Develop and maintain a website for a newly funded research project.Feature extraction for the image of the seismic wave graph using traditional computer vision algorithms such as HOG and more.

Developed the fake news classification model with traditional machine learning algorithms including XGBoost and Random Forest using Scikit-learn to achieve 85% accuracy in binary classificationIntroduced the Non-Negative Matrix Factorization on the top of traditional information retrieval TF-idf technique and increased the performance by 30%Achieved competitive result compare to the traditional methods by proposed and developed the LSTM neural network with attention mechanism in the downstream task by Pytorch

Built the data visualization dashboard and report generation tool with JQuery and D3.js and reduced the user daily workload by 95%Designed and implemented the RESTful API with third-party data source integration using Django and MySQL