Hello, I am Tim

About Me

Data Champion
Critical Thinker
Versatile Problem Solver
Fast Learner

LinkedIn: https://linkedin.com/in/timzai/

Projects

Creating and Evaluating Successful Cryptocurrency Exchange Strategies

https://mads-swaps.github.io/

A deep dive into several different strategies that can be used to predict buy and sell opportunities for the exchanging of cryptocurrencies. Utilizes Amazon Web Service serverless architecture patterns to automate and scale the machine learning models in a production environment. Supports a variety of classifiers and regressors from scikit-learn, ensemble of models, as well as neural networks.

time-series analysis market forecasting AWS Serverless Application Model cryptocurrency binance statistical arbitrage scikit-learn ensembles random forest logistic regression GradientBoost AdaBoost XGBoost ResNet multilayer perceptron

Improving Earth-orbiting object state prediction by using big data analysis and machine learning techniques on TLE data

https://cm-tle-pred.github.io/

Utilizing deep learning, and unsupervised learning on publicly available TLE data of Earth-orbiting objects to deliver superior orbit propagation results compared to the SGP4 standards.

time-series forecasting deep learning PyTorch neural networks signal processing event detection data reconstruction big data analysis space situation awareness orbit propagation non-resolved object characterization

Overwatch League Video Replay Teamfight Efficiency Analysis

https://timzai.github.io/owl-teamfight-efficiency/

Used computer vision techniques and deep learning to extract data from in-game footages to perform team performance efficiency analysis for the Overwatch League 2020 playoffs.

computer vision OpenCV Keras PyTesseract deep learning convolutional neural networks video sequence analysis time-series analysis signal processing information visualization esports Overwatch Overwatch League

Orbital Congestion

https://timzai.github.io/orbital-congestion/

Analyzed low Earth orbit collision risks by processing 150 million rows of satellite time series data. Highlighted as exemplar work and recommended for follow-up study and formal paper submission by space situational awareness and space traffic monitoring experts.

exploratory data analysis data structure efficiency time-series analysis rolling-window analysis signal processing outlier detection information visualization animated visualizations space situation awareness