Risk solutionLead accident reporting web platformProactive claim handling with machine learning models

Full stack developmentReact, Typescript, Python

Dropbox Paper teamFull stack developmentReact, Typescript, Java

This is a project of Digital Biology Lab.Built a Java-based standalone application to predict protein phosphorylation sites using Java:• Created machine learning models using KNN, SVM classifiers and Blast.• Added new features such as KNNScore and Grey Information and new PTM (Post-translational modifications) types. • Coded Cross Validation to test the accuracy of prediction after adding new features, which grows from 83% to 85%.

Developed and deployed the official website using HTML, CSS, PHP, JS, AJAX, JQuery.Build the control software of NovaTest P100 (a gas chromatography device) using UWP app using ASP.NET core in C#.Designed and implemented peak finding, integration calculation and noise reduction algorithms to analyze chromatography data. Optimized SNIP algorithm to make baseline detection of chromatography more accurate by 30%.Developed a program on Arduino Mega Board microprocessor for the digital and analog I/O and data gathering using C.

This is a project of Distributed and Intelligent Computing Lab.Built a web application to help staffs at Unilever (Jefferson) to manage equipment through a virtual map using VS, PhoneGap, html, CSS, PHP, JavaScript, JQuery and bar code recognition:• Implemented the modules of marking the machines’ locations and tracking the maintenance staff’s real-time locations based onGoogle Map.• Stored the records of maintenance of the machines to MySQL on AWS cloud.

Developed data analysis models in self-service financial transaction terminal project of Postal Savings Bank of China using C++, ClearEx, Flash MX, GHOST, etc.Implemented functions of sending, receiving, parsing and analyzing the 8583 protocol messages from ATM/POS.