An expense manager web-app to track and analyze individual and group expenses.

CoronaTimes is a simple single page web app where users can share their experiences dealing with the COVID-19 pandemic. Users can fill out a form and submit it, and can read the submission by other users. The prime goal of this personal project is to experiment with creating a form-auto-saver microservice, where the user's progress is auto-saved in a document store without the user's intervention.

Implemented a parallelized k-NN classifier algorithm for NIST handwriting recognition data. Applied Broadcast Join and 1-Bucket Partitioning using Hadoop MapReduce and Scala Spark.

A multi-threaded socket chat application with features like individual-chat, group-chat, message queueing for offlinemode, blocking abusive language, read receipts, end-to-end encryption, and multiple MIME type messages.

Achieved 89% accuracy in classifying the images of ships and icebergs using Convolutional NeuralNetworks (CNN) models on the StatOil/C-CORE satellite image data.

Achieved an accuracy of 96% in predicting the bankruptcy of ~5000 Polish companies; experimented withmultiple ML models like Logistic Regression, Decision Trees, Gradient Boosting, Random Forests, SVM, etc.

A Hackathon-finder single-page application where organizers can post the details of hackathons and participants can search and bookmark hackathons and interact with the organizers.

A website/blog maker web-app where users can create websites, add pages to websites, and customize the pagesand sites with widgets like Flickr images, embedded videos, rich text content, etc.

Analyzed Naïve Bayes and Support Vector Machine (SVM) classifiers for Sentiment Analysis for conversational texts. Pre-processed the data set and calculated the word frequency for each word in the corpus used for the Naïve Bayes approach. A lexicon-based approach is employed for SVM approach, by making use of SentiWordNet—a lexical resource for opinion mining.

Developed the prototype of vehicle parking monitoring system which comprises of an Android mobile app that can be used by employees of an organization to receive notifications of available parking slots for their vehicles. A different faceless Android-enabled device publishes the entry/exit events of vehicles when employees swipe their allotted RFID-tag against it, through Near Field Communication (NFC) protocol.

Developed a basic database IDE that can be used to work on a remote database. Used Java SE + Swings APIs to build the GUIs and JDBC for the database connectivity, with Remote Method Invocation (RMI) for the remote database access.