Prajakta Tathavadkar

Computer Science Graduate Student
Stony Brook University, NY

See my work Profile Picture

Education



December 2014

Master's in Computer Science

Stony Brook University, NY

GPA: 3.67

Relevant Courses:Analysis of Algorithms, Artificial Intelligence, Data Mining, Natural Language Processing, Operating Systems, Asynchronous Systems

June 2013

Bachelor's in Information Technology

Mumbai University
Distinction with 72.8%

Was amongst the top 2% of the department

Skill Set



Languages Java
Python
C
Basics of C++
HTML 5
CSS
JavaScript
JSP
Servlets
Basics of jQuery and AJAX

Databases MS SQL
MySQL
Oracle
MS Access
Sybase
IBM DB2
BM Netezza

Tools, Frameworks Eclipse
Visual Studio
PyCharm
.NET 3.5
Drupal
MS Office
GIT
Subversion

Operating Systems Linux
Windows

Work Experience



Jun 2014 - Aug 2014

Bank of America, Merrill Lynch

Software Development Intern

Integrated the AMPS message publishing framework into the Feed processing engine to produce information at real-time. This was done in an efficient manner by appropriate use of Data structures and minimized latency. Also, implemented a subscriber in Quartz Framework (Python based) to receive these messages. I collaborated with other teams to enable AMPS Browser compatibility making it easy to use.

Nov 2013 - Jan 2014

CA Technologies

Research Aide

Developed a listener to monitor the real-time Audit data of different clusters. This was done in a multi-threaded fashion for maintaining transparency amongst various clusters. Also got acquainted with Hadoop environment setup and worked with sample Map Reduce jobs on the same.

May 2012 - Jun 2013

Omnitech InfoSolutions Pvt. Ltd.

Software Development Intern

Worked with 2 of my classmates to build a Cloud Based Multi-tenancy Spares Management system. As a software developer on this project, I developed modules for multi-tenancy and metering functionality.

Projects



HearSay

Hearsay is a web browser tool for the visually disabled people. It aims at building a web browser plugin to navigate through the web pages via Audio and consists of Java and JavaScript components. I was responsible to process the mutations in HTML DOM in the JavaScript to enable dynamic web page navigation. Currently, I am working in a team of 2 to build a componentized framework which could be migrated to the cloud.

Personal Analytics on Movie Characters

Developed a new classification algorithm to detect 9 personal traits of a movie character (positivity, etc.) based on its dialogues. Improved the efficiency by 4-5% from the baseline of Unigram and Bigram using POS tagging, Function words (‘my’, ‘the’, etc.) and punctuations, which we identified as potential trait classifiers.

Operating Systems

Developed a new classification algorithm to detect 9 personal traits of a movie character (positivity, etc.) based on its dialogues. Improved the efficiency by 4-5% from the baseline of Unigram and Bigram using POS tagging, Function words (‘my’, ‘the’, etc.) and punctuations, which we identified as potential trait classifiers.

Fake Review Detection

Implemented a system for the detection of fake reviews for Amazon products by applying feature, timeline and graph based probabilistic data mining techniques to process around 2 million entries of data.

Memorable but Incommunicable Passwords

Implemented an image based password protection system against phishing attacks. Was responsible for developing the website as well as managing the database while my team mate handled the designing and surveys of the system.

Pac Man and Optical Character/ Face Recognition

Improved the efficiency of Pac Man by applying DFS, BFS, pruning, reinforcement learning, etc. Also applied some Machine Learning concepts like Bayesian learning, etc. for recognizing textual facial data.

E-Commerce Website

Built an online shopping website with 3 of my team mates for IBM’s “The Great Mind Challenge 2011” competition. As a developer, implemented modules such as shopping cart, advertisements, etc.

Fraud Detection

Applied K-means clustering and prediction techniques on sample data to detect anomalies in the same. The algorithm was optimized and implemented to enable scalability for large datasets.

Chain Replication Algorithm

Currently implementing the famous chain replication algorithm for supporting high throughput and availabilty as a part of the course Asynchronous Systems.

Contact Me