Siddhesh Dilip Rumde is a graduate student at the University of Southern California pursuing a Master's degree in Computer Science. He has experience developing applications using technologies like Java, C++, Python, PHP, Android and databases like MySQL. Some of his projects include developing classifiers for emotion detection in tweets using Naive Bayes and SVM algorithms and creating a real estate property search application using APIs, PHP, AJAX and porting it to Android. He has also worked on developing a GUI for an iRobot using VB.Net and an Android application for sign language translation.
1. Siddhesh Dilip Rumde
2707 Portland St., Unit #311, Los Angeles, California 90007| (323)283-2416 | rumde@usc.edu
Education:
Master of Science, Computer Science (Fall 2014) Expected May 2016
University of Southern California, Los Angeles, California
Current GPA: 3.74/4
Bachelor of Computer Engineering June 2014
Sardar Patel Institute of Technology (SPIT), Mumbai University
Overall Aggregate: 71.27%.
Technical Abilities:
Programming Languages: Java, C++, C, VB.Net, Android Programming (Java), Python
Database Technologies: MySQL, MS Access, SQLite, PostgreSQL
Web Technologies: HTML, CSS, JavaScript, PHP, XML, JSON
Development Tools: Netbeans, Eclipse, Microsoft Visual Studio, Android Studio, Adobe Dreamweaver
Operating Systems: Windows (7/8/8.1), UNIX, Linux (Ubuntu, Fedora)
Work Experience:
Research Assistant, Sardar Patel Institute of Technology (SPIT), Mumbai University Aug. 2013-May 2014
Worked under Dr Dhananjay Kalbande as a research assistant to work on the project “Sign Language
Translation for a Mobile Application”.
Database Designer, NCC Limited, India May 2013-Aug. 2013
Assembled the database for the construction company so that there is optimal storage of the assets managed
by that company.
Projects:
Multi-Label Classification Approach for Emotion Detection in Tweets Mar. 2015-Apr. 2015
Developed classifiers for individual emotion detection after filtering the tweets by removing unnecessary
information such as URL, handle name which have no relevance with the emotions in the tweets. Implemented
classifiers using Naïve Bayes and Averaged Perceptron Technique along with the use of SVM.
Generated features for each emotion by using a combination of unigrams, bigrams, trigrams, synonym sets,
punctuations and topic modelling using Mallet. Achieved an F-Score in the range of 0.6-0.75 for each emotion
that we classified.
Parts of Speech Tagger (POS) and Named Entity Recognizer (NER) using Averaged Perceptron Feb. 2015-Mar. 2015
Implemented a generic multiclass average perceptron in Python for classification of data and use the model
thereby generated to tag a word's Part of Speech and to recognize its Named Entity.
Achieved F-Score of 0.95 on POS and 0.71 on NER for around half a million words
Spam and Sentiment Classification using Naïve Bayes Technique Jan. 2015-Feb. 2015
Developed a generic spam and sentiment text classifier in Python using Naive Bayes classification to classify a
message in different classes by training the model with more than 18,000 training files and testing it on 1500
test files for Spam Dataset and 25,000 training files and 2000 test files for Sentiment Dataset.
Achieved approximate F-Score of 0.95 on Spam Dataset and 0.84 on sentiment Dataset.
Real Estate Property Search Application Oct. 2014-Dec. 2014
Developed a web and android application using Zillow APIs. Created a responsive web UI using Bootstrap
components. Made use of AJAX calls to the PHP file located on AWS Elastic Beanstalk Cloud. Retrieved JSON
format data from XML data which was later parsed at the client side. Ported this to Android mobile device.
Technologies Used: PHP, AJAX, CSS, HTML, JavaScript, jQuery, XML, JSON, Android SDK, Facebook SDK.
Mission Science iRobot Sept. 2014-Dec. 2014
Worked on a Soft Engineering Project to develop a GUI for the Mission Science iRobot. Used Incremental
Commitment Spiral Model for the development of the GUI.
Technologies Used: VB.Net, WinAVR.
Android Application for Sign Language Translation Aug. 2013-May 2014
Implemented an Android Application which captured the signs made by the user using the device camera and
translated it to the corresponding word. Used contour analysis by correlation methods to obtain the results.
Technologies used: Java, Android SDK, Android NDK, XML.