Yunlin Qi

The fun and classic tetris game! With multiple levels of difficulties to choose for you to either relax a little bit or challenge your skills!

A powerful Autodesk App developed in Python that connects 3D building information modeling to 3D building analysis of steel members with just one click.

Java (Spring Boot) web development of a streaming event platform that dynamically tracks the status of various parties including studios, streaming service providers, streams, and demographic groups. Developed the system with archivability, scalability, and performance monitoring.

Python (Flask) web development of a dog shelter administration platform. Developed with comprehensive functionalities including multi-user login, dog registation & adoption, dog info management, expense management and monthly report lookup.

Android mobile app development of a simple word find game that allows a player to solve word puzzle, adjust game settings, and view statistics.

Application of logistic regression and neural network models in prediction of Covid ICU and hospitalization, based on 563201 records of patients' preconditions. Developed web app to visualize data distribution, model results, and interactive prediction test.

Application of Apache Spark in big data and cloud computing via AWS and GCP. Data processing and analysis on over one billion individual taxi trips from the NYC Taix & Limousine Commission.

Fun exploration and data visualization for various board games' ratings, popularity, and metadata, by utilizing Taleau, D3.js and HTML. Created force-directed graph showing similarities among games, as well as the interactive line-bar chart and choropleth map for more complex game data visualizations.

Application of various Transformer models in prediction of Grade 3-12 reading material readability based on 2834 entries. Performed model optimization and hyperparameter tuning to improve prediction accuracy.

Application of various VGG16-based models in digit detection and recognition of street view images. Data processing and machine learning carried out on 99289 street view images to train and optimize model.

The intersting implementation of pedestrian tracking for image sequences and videos. Applied the Kalman and Particle Filters algorithms on pedestrian tracking under various scenarios, such as single object, multiple objects, videos with noises, occlusions, and moving cameras.

Effective and cool content-aware image resizing for both image reduction and expansion. Implemented based on the seam carving algorithms presented by Shai, et al and Michael, et al.

papers here and here