Developed a web application using four loosely coupled microservices with their own load balancers and independent execution environment to predict future stock trend values and display the past and future stock trends on the dashboard using nodeJs

Optimized support vector regression (SVR) and linear regression (LR) prediction models to predict and determine next 30 days stock trends by extracting multiple features from multi-dimensional stock data provided by Yahoo Finance to obtain a confidence score of 0.929

Dockerized and deployed the application on Google Kubernetes Engine (GKE) and provided the orchestration features like Horizontal Auto-scaling and Load Balancing to increase the efficiency by 90% and achieved lower resource utilization

Technologies - NodeJS 13.8, Python 3.7, Google Kubernetes Engine (GKE), Docker 19.03.13, MongoDB 4.2.3, Bootstrap 4

Designed and Developed a web spider in python that crawls the website data into text using a Beautiful Soup python library and extracts useful data from HTML tags to help construct a meaningful dataset for the detection of phishing websites

Integrated whois lookup API and gathered URL metadata such as domain ownership and registration details, IP addresses, ranks, etc. for generating a rich dataset

Images and code not attached for organizational confidentiality purposes.

Technologies - Python 3.8, Beautiful Soup, whois, pandas

Constructed a model to determine the success rate of the movies (positive, negative, neutral) based on the classification of emotions within text reviews of movies on IMDB website using classifiers such as Logistic Regression, Support Vector Machine, and Naive Bayes

Preprocessed the dataset by removing HTML tags, Lemmatization, filtering stop words and special characters, and Text Tokenization

Feature Extraction using the Bag-of-Words approach which uses Count-Vectorizer that turns arbitrary text into fixed-length vectors by counting the frequency of each word

Compared the accuracy of classifiers and discovered that Logistic Regression outperformed SVM and Naïve Bayes

Technologies - Python 3.8, sklearn, NLTK, gensim

View Code View Report

Implemented a secure connection between client and server using TLS 1.2 protocol using OpenSSL in C wherein a client sends the request to the server that contains cryptographic data, the server then verifies the data and sends a proper response to the client

Generated the server's customized digital certificate for the mutual authentication that is verified by the client which provides identification, authentication, confidentiality, and integrity to the established connection

View Code

Implemented the substitution cipher method called Monoalphabetic cipher in C++ to encrypt the given plaintext string into a ciphertext for secure transmission and decrypted the generated ciphertext back into plaintext using an autogenerated key

Program (both encryption and decryption) prints the mappings generated between the plaintext letters and the ciphertext letters. The mappings are printed using the format a-c1, b-c2, …, z-c26, where ci is the corresponding cyphertext letter, and c1\=c2\=…\=c26

View Code

Developed a web application blogging system in nodeJs where users can register themselves, can edit/ update their account information, post or edit articles, and comment on any article

Designed user interface using the javaScript’s Logic-Less template library called MustacheJS and business logic in backend and error handling using NodeJS

Implemented REST APIs using ExpressJS that acts as an interface with the functionality of the backend business logic and helps retrieve or change the data from MongoDB database

Technologies - MongoDB, JavaScript, Mustache JS, express JS APIs, JQuery, HTML5 and CSS3

View Code

Implemented a software chatbot named Dexter in python that can engage in conversation with users

Incorporated a corpus of questions & answers that trains the bot using machine learning algorithms

Designed the interactive UI using kivy library in python that allows users to input text via keyboard or voice input via microphone

Images and code not attached for organizational confidentiality purposes

Technologies - Python 2.7, Kivy, Google Speech Recognition, NLTK

To know more about similar Chatbot - refer this link

Trained the system to classify the emails into spam or ham from the testing dataset using various techniques such as single and multi-layer perceptron, SVM’s, and probabilistic based classifiers such as Decision tree and Naïve Bayes

Data Preprocessing using tokenization, vectorization, language translation, and filtering stopwords

Compared and determined the accuracy of classifiers and discovered that SVM outperformed other two classifiers

Technologies - Python 3.8, Naive Bayes, Support Vector Machines, Perceptrons and Neural Networks

View Perceptron Code View Naïve Bayes Code

Developed a secure password management system using a symmetric encryption algorithm called Advanced Encryption Standard (AES) in Java that encrypts the password entered by the user and decrypts and verifies it if the password is correct using an autogenerated key

View Code

Implemented the Decision Tree Algorithm to determine the next attribute using two heuristics - Information Gain heuristic and Variance Impurity heuristic.

View Code

Developed a web application in Java Servlet and JSP having functionalities such as authorizing Donor & Organization, login donor & organization, search donors by blood group, display blood availability by blood group, and connect the donor and organization

Designed a highly interactive user-friendly website using HTML5, and Bootstrap3 template, data stored in Oracle 11g database and deployed using Apache Tomcat Server 7

Technologies - Java Servlet, JSP, HTML5, Bootstrap3, Apache Tomcat 7, Oracle 11g DB

View Code

Developed a Web-application in Spring Boot and MVC architecture providing all the internet banking functionalities such as register user, login user, credit, deposit, view transaction details, banking statements, and stored data in SQL database and deployed using Tomcat 7 server

Implemented user sessions using spring boot sessions to keep track of the customer behavior and store session data for a specific user

Technologies - Java Spring Boot MVC, HTML5, Oracle 11g DB, Apache Tomcat server 7

View Code

Built a hardware device that captures Human voice via microphone and subsequently processes these requests using AI techniques to respond through the speaker

Also developed an Android application along with IOT box consisting of LCD screen, Curtain control & Lights control and demonstrated the interconnectivity between these devices and how they work in a synchronized manner. IoT box consisting of 3 modules – Drapes control, Mini LED screen, and Neo-pixels lights to show how we can turn on/off these modules just by our voice commands or by an android application.

Integrated google firebase for authentication and user authorization and synchronized assistant, android app, and IOT box by storing user requests using a firebase server

Technologies - Python 2.7, Raspberry Pi, C++, Arduino Uno, Java, Android, Speech Recognition, NLTK, Natural Language Processing

Citation - Abhay Dekate, Chaitanya Kulkarni, Rohan Killedar "Study of Voice Controlled Personal Assistant Device". International Journal of Computer Trends and Technology (IJCTT) V42(1):42-46, December 2016. ISSN:2231-2803. www.ijcttjournal.org. Published by Seventh Sense Research Group.

View Journal