Ziyu Cheng

A Directed Research Project to visualize workflow and team relation in software development process.
· Designed data structures and RESTful APIs, and created front-end using JavaScript.
· Designed and realized customized chord diagram to make visualization of team relationships.

Integrated Autodesk Forge Viewer in USC’s Facilities Management System.
Developed core features in Viewer, including navigation, category, system hierarchy, etc.
Generated Revit plugin to export MEP System Graph Structure to provide extra customized model data.

Designed server-side conceptual model using UML.
Took charge of database design in MySQL.
Developed backend server based on Spring + Struts + MyBatis framework using Java.
Generated frontend using JavaScript with Jquery, Ext.js.

A mobile card game based on Unity3D.
Worked with 2 front-end engineers, 2 back-end engineers, 2 designers and 1 game graphic designer, developed a cross platform mobile game with 3D scene, LBS support and chat function.
Developed several modules in client end using C#.

Operation Management System for "Kong Fu" mobile game.
Developed key features of back-end server using Java, including operation on account management, authorization management, start/stop sever, GM operations, calculation and visualization of statistic data(DAU, WAU, MAU, DRR, WRR, DRR).

A intern development project with NeuSoft in the summer of 2011.
Developed a graphical interfaces client from scratch using Java to simulate mobile phones and a server system interact with GUI client to simulate message and payment service.

A Student Research Training Program in University of Science & Technology Beijing.
Developed android client using Java.
Developed barcode information server using Java.

An international Machine Learning Competition of match community raised questions with domain expertsheld by IEEE-China and Toutiao.
Implemented preprocessed data and generated features, implemented and applied PCA on 100 GB data from scratch usingPython in Google Cloud Linux Environment.
Improved training speed over 15 times by introducing parallel process using Python.

Decision network capable of calculation with uncertainty and observed events written in Python.
Implemented key algorithms including enumeration, elimination, maximum expect utility and
decision making algorithms.

An Operating Systems course Project, a mini Unix-like OS kernel written in C in Linux Environment.
· Implemented core components including process management, drivers for terminals and hard disks, VFS, and page-based virtual memory.
· Implemented key system calls including fork, waitpid, execve, open, read/write, mmap, sbrk, etc.
· Completed capability of loading and running gcc-compiled programs