Real-time Spark Streaming Analysis & Sentimental Analysis on Meetup Streaming data.

• Successfully build Lambda Architecture based real time streaming analysis, sentimental analysis system deployed on Amazon AWS
• Analyzed streaming data from meetup based on RSVP, Events, Events comments to provide real-time updates of number of events trending based on State, Country visualized them on google maps.
• Sentimental analysis on comments to classify event comments into good, neutral, bad which helps event organizers and users to judge event.
• Technologies: Spark Streaming, Kafka, HBase, Hive, NodeJS, MapR, Mapr Hbase Rest, Google Maps

• Used Yelp Business, User, Check-in, Review, tips data set to provide classified business and user recommendation.
• Implemented business establishment recommendation based on type of business, location and local competitors with Jaccard Distance, collaborative filtering and regression. User based recommendation based on KNN classifier algorithm.
• Technologies: Java Spring MVC, HTML5, Material Design, AJAX, D3 JS, Rickshaw Charts, Google Maps

Zeus is a platform as a service(PAAS).

- Users can write, compile and deploy Nodejs applications in DevOps software development method. - Users are given a sub-domain on deploying their applications successfully. - Responsible for storing User data and retrieve from MongoDB using RestAPI. - Also in implementing GitHub Clone API. - Technologies used: Express.js, Nodejs, JQuery, Ajax, MongoLab Developed a multi-cluster NodeJS based web hosting application. Implemented features such as creating free sub-domain per application, a basic template for a user to work upon, real time online compilation, listing directories and sub directories recursively. • NodeJS, REST web services, Amazon AWS (Route53, EC2), MongoDB, EJS.

Goal of the project is to build highly available, consistent and scalable distributed system. • A communication and storage system to collect, store and stream images. • Image messages sent by user (client) are broadcasted to all clients connected to local cluster as well as different clusters. • This project focuses on communication performance, overlay network flexibility, leader election, and voting. Implemented Raft consensus algorithm in dynamically connected mesh topology in core netty server. • Multithreaded Java application with server implemented using Netty (Java NIO) and Protobuff for messaging. • Technology Stack: o Network Server: Core Netty o Message Passing: Google Protobuf o DB Layer: FTP Server o Server Side Coding Language: Java o Client Side Coding Language: Java

The project goal was to create a multitenant web application for different Software Methodologies  Waterfall -> Gantter  Scrum -> Easybacklog  Scrum/Agile -> Kanban • Data designed for multiple tenants and users in MYSQL. • Tenant id used to implement multitenancy within different tenants. • Openshift was used to host node.js and MySQL servers. • Cloud9 IDE was used to collaborate in team. • Data access exposed using REST Webservices. • HTML5 and Twitter Bootstrap used to construct UI web pages. Technology Stack • Node.js • MySQL • AJAX • JQUERY • HTML5 • BOOTSTRAP • OPENSHIFT

This iOS application uses facebook to authenticate the user, and create / load the record of the users stored in our database for user’s session. The application upload any image provided by user to our server on EC2, which eventually upload to S3 Simple Storage Solution. Application will fetch images from S3. Our server will write and read to/from Nosql database in the cloud (MongoDB). Major features include: - Login with Facebook / or as guest . No registration required - Create an album - Browse albums and photos - Upload photos into an album with photo metadata, location and set to private or public - Share photos with your friends via email invite - Search by keywords (will match with name, location or metadata - Guest user can search for all public photos only. - Registered user can search for their albums and photos also.

This project involved simulation of Student Union(cafeteria) of SJSU in code over a distributed environment using Python. The requirement was to have a highly scalable architecture and design, which will allow the user to add multiple servers, chefs, cashiers & customers. Technologies Used: Python, Flask

Google Plus Mini App allows user to view his/her profile, circles, friends by integrating with Google Plus (G+) Domain APIs.