I am a graduate student at the University of South Florida, completed my Master's degree in Computer Science. A passionate software developer looking for full-time opportunities where I can apply my problem-solving skills and knowledge in Computer Science to solve challenging real-world problems.
-
Experience
Research Assistant, CIBER Lab, University of South Florida January 2019 – December 2019
• Developed Java and Python scripts for the research team as per their requirements in the experiments
• Created app usage-based images using LabelEncoder from a customized dataset containing app usage logs of users, also generating additional images using Gaussian filters
• Designed a deep neural network using Keras for classification of about 300,000 generated user images, which achieved an F-score of 96.8%
Publication: A Closer Look at Mobile App Usage as a Persistent Biometric: A Small Case Study – https://arxiv.org/abs/1912.11721v1
-
Projects
1.) Recommender System using Netflix Prize Dataset
• Developed a recommender system which achieved an accuracy of 95.3% using Pearson’s Coefficient
• Evaluated and compared the performance of different collaborative and model-based filtering techniques using 5x2 fold cross-validation and RMSE values
Technologies Used: Python
2.) Profiling Internet Users
• Developed an application to observe and analyze the usage trends in internet from the Cisco NetFlow (version 5) data logs
• Used Spearman’s Coefficient to compute the P-values for all combinations of users to distinguish individual users
• Achieved at least 60% better performance by using the ‘sqldf’ module to optimize code in python
Technologies Used: Python, SQL
3.) Real-Time Classification of Tweets
• Created a Web Application to classify real-time tweets from Twitter API, which is at least 10% faster than existing systems
• Implemented three modules to offer user-friendly service with an intuitive interface for the overlapping keyword, location, and retweet search filters to classify tweets and generate custom results
Technologies Used: Java, Apache Tomcat, HDFS, Twitter API, JSP, JavaScript, Spring
4.) Image Recognition using AWS Rekognition
• Used AWS Rekognition, boto3 module and Pillow in python for recognition of objects in images
• Returns the copy of the supplied image as a Pillow image which contains the name tag of the object that is identified
Technologies Used: Python, AWS Rekognition, boto3, Pillow, Requests, AWS IAM
5.) Substituting Buffer Replacement Policy in PostgreSQL
• Changed the existing clock buffer replacement policy and replaced it with least recently used (LRU) buffer replacement policy
• Used FIFO queues for storing and retrieving buffers and modified underlying functions in various files of the original software
Technologies Used: C
6.) Real-Time Operating System
• Implemented a real time operating system in a layered approach, including designing a bootloader using assembly language
• Added functionalities like multitasking by designing a trap frame handler, system calls and exception handler
• Applied Copy-On-Write functionality and process page sharing mechanism using C language
Technologies Used: .asm, C