Resume

Sign in

Python, Java, C, C++, PHP, Perl, Machine Learning, Data Science

Location:
Chicago, Illinois, United States
Posted:
April 22, 2019

Contact this candidate

Resume:

ANH TUAN TRAN

Phone: 208-***-**** Email: ac86fq@r.postjobfree.com

EDUCATION:

The University of Illinois at Chicago, Chicago, Illinois GPA: 3.2 Master of Science in Computer Science

Anticipated Graduation Date: May 2019

Boise State University, Boise, Idaho GPA: 3.8

Bachelor of Science in Computer Science

Graduation Date: May 2015

SKILLS:

Programming Languages: Python, PHP, HTML, JavaScript, CSS, Python, CUDA, Perl, Java, C++, C

Experiences: Open Scene Grap, WebGL, Object-oriented design patterns, Relational databases experience with MySQL, MSSQL and SQLite, Schema Design. Framework: ASP.NET, Hadoop, MPI, NodeJS, D3, Jupyter Notebook EXPERIENCE:

Software Engineer

Test Department, Micron Inc, Boise, Idaho

June 2015- June 2017

● Developed software tools and web pages that display data, visualization about different testing processes using Perl and SQLite database

● Worked in team of three to develop a web site for releasing testing software using PHP, JavaScript, JQuery, SQLite, and Python

Visualization of EPSCOR data on tiled display/Internship Computer Science Department, Boise State University, Boise, Idaho Feb 2014- May 2015

● Worked in team of 2 under Dr. Steve Cutchin’s supervision

● Built a stereo display machine for high resolution images using Open Scene Graph. Integrated Oculus SDK into the machine so that an Oculus can be used with the machine

● Integrated the machine with web browsers using WebGL so users with an Oculus can use it via the Internet

ACADEMIC PROJECTS:

Senior Lifestyle Analysis

April 2019

The project was to analyze data from Senior Lifestyle Corporation. The data includes 4 datasets about prospect demographics, sale activities, resident move in/move out dates and revenue data. Jupyter Notebook was used to as the analysis tool

● Determined the best combination of activities to get maximize conversions and reduce labor using ExtraTreesClassifier. Built a SVM classifier using those activities to predict if a prospect can be converted to a resident and used 10 fold cross validation to evaluate the classifier with the accuracy was 95%

● Grouped prospects into 3 subgroups based on the how quickly they were converted into residents and determined the attributes can be used to categorize them. Built a SVM classifier to classify prospects into those subgroup and used 10 fold cross validation to evaluate the classifier. the accuracy was 97%

ReDos Attacks’ Monetary Analysis on AWS

April 2019

The project was to analyze how fast ReDos attacks drain your money on serverless computing platform. AWS Lambda was our chosen platform.

● Created a Lambda function in NodeJS with/without a ReDos vulnerable call on AWS Lambda and used jmeter to create calls to it to investigate how much a ReDos attacks consume resources in comparison to a normal call. The rate increases with the increase of the timeout limit

● Created Python program that can run sequentially and in parallel. When the memory capacity is set under 1.8GB, AWS Lambda container executes everything sequentially. For the capacity bigger than 1.8 GB, the runtime is faster which suggests parallel execution but only fast enough to support the hypothesis the CPU is hyper threading Visualization of Longitudinal Variability of Usage of Cigarettes and e-Cigarettes December 2018

Consultant: Dr. Robin Mermelstein, UIC Institute of Health Research and Policy I worked in team of three to create a story telling and analysis tool to help Dr. Mermelstein to understand longitudinal usage and mood changes of individuals who smoke cigarettes and e-cigarettes over time by variability of different users

● The visualization tool was web-based and created with NodeJS

● We discussed with Dr. Mermelstein to build different graphs to visualize the usage of cigarettes and e-cigarettes and affected mood changes over time using D3 Tweet Sentiment Analysis

May 2018

The project was to classify tweets’ sentiment into 3 categories: -1 (negative), 0 (neutral) and 1 (positive). The tweets were about Obama and Romney during their 2012 presidential campaigns.

● Applied text cleaning and pre-processing techniques (tokenization, n-grams, tf-idf vectorization, etc.) to use as input for machine learning models

● Built different machine learning models using Naive-bayes, Support Vector Machine and Neural Net. The LSTM Neural Net built with Keras framework had the highest F-score of 0.68



Contact this candidate