Compute Infra (Ray) for Machine Learning Training, Evaluation, and InferenceParallel Machine Learning Training Platform

MSAI of Office 365, ML Inference platform, Model Deployment, Model Serving

Data Engineering, Apple Card project (https://www.apple.com/apple-card)

Infra Planning Systems

• Initiated and Implementing capacity portal project that drives capacity planning & forecasting of Salesforce.com• Demoed capacity portal POC and been selected as Best Technical Member of the Month• Building the capacity tools using Java, Web Services, jQuery, Bootstrap, HBase, OpenTSDB, etc.• Delivered the infrastructure features that drive migration of customer data

[Machine Learning for Search Trending Prediction]• Initiated and implemented a trending topic prediction engine by using web query data and majority voting machine learning algorithm to predict trending topics with high correction rate and prediction time in advance.• Scaled up the prediction engine by migrating it to grid computing environment with Hadoop by using Mappers to increase the computational efficiency.• Presented this intern project to CEO, and won the 1st place award among all internship projects within department

Focus Areas: High Performance Scientific Computing / Statistics & Mathematics for Big Data

[Machine Learning for Internet Security]• Created a computational engine to compute IP reputation scores by using machine learning technique (logistic regression classification and ensemble learning) to improve the correction rate of intrusion detection up to 90% in large security system project hosted by X-Force research (Atlanta, GA) • Presented the IP reputation project to customers and large enterprises (such as NTT West in Japan)[Cloud Computing (SaaS)]• Built SaaS cloud services using MVC model to simplify development process to nearly 30% on the IBM cloud computing platform for its developers• Created a hybrid cross-platform mobile web application for China Ningbo logistics companies to ensure their on-time shipment checking[Engagement of Customer Projects]• Represented as speakers in IBM Rational Developer Forum to introduce IBM Rational Team Concert & Cloud Computing products to the customers and large enterprises• Developed a the proof of concept of artifact management feature for enterprise asset management system for a automobile manufacturer of Germany (BMW) using Java and Eclipse• Created a hybrid cross-platform mobile web application for China Ningbo logistics companies to ensure their on-time shipment checking

[Mobile Application Development]• Developed mobile applications on Windows Mobile for and collaborated with telecom operators in U.S. (Verizon and Sprint), and China (China Mobile) • Collaborated with Microsoft Windows Mobile Application developers to customize telecommunication applications for AT&T, Verizon and Sprint