Full Stack Data Science

July 2018

• Built a full stack data science web application at the Viasat Intern Hackathon using Django and PostgreSQL, to increase customer engagement by prioritizing and categorizing customer reviews in real-time

• Preprocessed the raw data to implement Doc2Vec algorithm and an SVM classifier for the machine learning model
Machine Learning

BU Autonomous Race Car

December, 2017 - May 2018

• Worked with Dr. Renato Mancuso on developing an Autonomous Race Car with an objective to train a model that can provide coarse grained localization without GPS, to compensate mainly for loss of localization or to detect spoofing

• Examined different approaches to develop new algorithms for Computer Vision involved in Autonomous Vehicles, to address the current safety concerns
Computer Vision
Machine Learning
Embedded Systems

Fake News Detection

February, 2018 - May, 2018

Developed a machine learning program to identify unreliable news based on its content. Achieved an accuracy of 94.53% using a Long Short-Term Memory (LSTM) model.
Natural Language Processing
Machine Learning

Big Data Containers

February, 2018 - April, 2018

• Built an Open Service Broker for the Dataverse API on the Massachusetts Open Cloud (MOC) to enable Big Data Analytics applications on OpenShift environment to consume data from Dataverse. Collaborated with people from RedHat, MOC and the Dataverse team at Harvard University to achieve the goals of this ambitious project

• What our service broker is essentially attempting to replace is the world of opening a service ticket for someone to send you a user ID and password with which you can connect to some system and then you put those user ID and password into some config file that will let your application connect. That's the old school way of connecting your application to a system and our service broker is trying to provide a standardized way to do that

• In our case, we have taken the OpenShift and the Dataverse and provided a connection between the two that helps automate the process I described above

Object Recognition

February, 2018 - March, 2018

Programmed an algorithm using a Faster R-CNN model to successfully identify and locate 90% of the objects in a given image, without sacrificing the speed of detection.
Computer Vision
Machine Learning

Network Visualization for Big Data

February, 2018

• Built a web application using JavaScript, HTML5 and CSS for better visualizing, managing and analyzing a complex network of nodes within a large dataset

• Came in 2nd Place at MIT CAVE Lab Hackathon 2018. Awarded for the originality and practicality of the application

3-D Gesture Controlled Game Against AI

November, 2017

• Developed a 3-D gesture controlled game in 12 hours at HackWITus 2017 using Unity Game Engine and Myo armband

• Configured the Myo armband to move and attack the AI bots using different hand gestures

• Programmed the AI bots to constantly search for the player to attack using Predictive Modeling and Pathfinding Algorithms
Unity Engine
Artificial Intelligence
Myo Armband


October, 2017 - November, 2017

• Developed an Android application for social networking using Google’s Firebase and Android Studio, to address the need for a new platform for local events and advertisements

• Used Open Data to display valuable information to the users about the neighborhood. Sorted the events according to the preference of the users for better user retention
Android Studio

Face Recognition and Verification Software

October, 2017

• Developed a web application at BostonHacks Fall 2017, which successfully detects and recognizes a person's face and displays information about the person from the database

• Integrated the OpenCV library in the back-end of our website which detects the face using Haar Cascade Classifier, and recognizes the face using the Eigenfaces Algorithm. The website was developed using HTML, Django and MySQL

Patient Monitoring System

August, 2016 - February, 2017

• Implemented a project which continuously monitors a patient-at-home’s pulse rate and temperature, and alerts the nearby emergency services using the GSM module

• Implemented an RFID reader which would read the unique ID stored in each RFID card that is given to every patient

• Developed a website using PHP, CSS and MySQL for ease in maintaining and updating a patient’s medical records
Atmega8 Microcontroller

Wireless Sensor Network with IoT Logging

June 2016 - July 2016

• Implemented a sensor node that could sense data from any connected sensor and transmit it over a WiFi interface to a Sparkfun Phant Server using ESP8266

• Storing values on IoT Cloud Phant using ESP8266, which are received by WSN co-ordinator from multiple discrete nodes using NRF24L01P
Embedded Systems

USB and Wireless Gesture Controlled Mouse

June 2016 - July 2016

• Used the on-board accelerometer (Freescale MMA8451Q) and capacitive sensor APIs of MBED and developed a mouse that alters its sensitivity with respect to the user input

• Extended the gesture controlled USB mouse to give it a wireless capability using an nRF24L01P trans-receiver pair
Embedded Systems

Multinode Wireless Data Acquisition System

June, 2016

Developed a data acquisition and logging system using multiple sensor nodes and a single receiver which logs the data. Communication was achieved using an nRF24L01P trans-receiver pair.
Embedded Systems

Custom Home Automation Controller using PSoC Bluetooth Low Energy (BLE)

July, 2016

Using Cypress Programmable System-on-Chip, developed a custom BLE profile to control various digital appliances using Bluetooth Low Energy.

High Speed Motion Detection

May, 2016

Detection of fast and abrupt motion of a device using a 3-Axis Freescale Semiconductors MMA8451Q accelerometer with FRDM KL25Z.

PSoC Bluetooth Low Energy (BLE) FindMe Beacon

May, 2016

Using the Cypress Programmable System-on-chip, developed a beacon with different alert modes to locate the PSoC 4 BLE device.

Digital Calculator

February, 2016 - April, 2016

• Designed a calculator using microcontroller IC 8051 that performed basic arithmetic operations like addition, subtraction, multiplication, division and squaring

• Studied and implemented the interfacing of various peripherals such as 16x2 LCD Display and 4x3 Matrix Keypad with 8051 microcontroller
IC 8051

Metal Detector

July, 2015 - September, 2015

Developed a basic metal detector using a 555 Timer IC, which successfully detects a metal and generates sound with the help of a speaker.
IC 555