Research Projects
1. DataPath System
It is cache-conscious, data-centric (rather
than compute-centric) database. I worked on its query optimizer.
For more information visit:
2. Online Aggregation for Large MapReduce Jobs
Read abstract of the Sigmod paper
Internship Projects
1. Table Analysis Tools (TAT) for Cloud (Summer 2008 Internship
Project @ SQL Server Data Mining)
TAT Cloud is a set of canned data mining tasks that you can use without having
SQL Server installed on your machine. It consists of encapsulations of some
common data mining problems, such as detecting key influencers, forecasting,
generating predictive scorecards or doing market basket analysis. The tasks can
be executed from
browser as
well as Excel 2007 (after installing TAT add-in).
If you had attended KDD 2008, you must have got opportunity to see
its demo (given by Raman and
Bogdan).
2. SpokenWeb (Summer 2011 Internship
Project @ IBM Research Lab)
In-progress
3. Embedded Web Server using VxWorks Real Time Operating System (@ ECIL Hyderabad as part of PG Diploma)
Acts as a standalone web server with remote file system. Since it is booted via
RS-232 (serial port), it does not require a hard disk.
Download
4. Sure Serve (Server Monitoring Utility) (BE Final year Project @ Rediff)
Allows the administrator to monitor server performance based on the specified
parameters. It comprises of modules (TCP, HTTP, Database and Application) that
monitors major functional areas of a commonplace web server. It plots the
parameter at real time using Multi Router Traffic Grapher.
Industry Projects
1. Usage Reporting of Hotmail data (Data warehouse)
Gathers data directly from product teams, transform and load into data warehouse
for aggregation, and generates reports for partners.
2. ERM (Employee Resource Management) website
An Ajax based web application by the means of which MAQSoftware manages its
employee timesheet details, project resource allocation and report generation.
3. Crystal (for Swedish Sleep Institute)
To improve existing 14 legacy applications and migrate them to ASP.NET and SQL
Server (so that they are accessible and housed under a single dashboard with a
single sign-on).
Course Projects
1. Single-user database
Implemented a simple database (using C++) that performs basic inserts (low-level
read/write), indexing (using B+ tree) and query processing. It conforms to TPC-H
standard and works efficiently on 1 GB data load. However, issues such as
concurrency control, failure recovery, etc are not addressed.
2.
Group Communication using UDP
Implemented a group management protocol using multicast udp messages. The
protocol will allow individuals to join a particular group based on simple
majority voting and will also deal with members leaving the group. The leader
election algorithm is a simple "bully algorithm". For futher details, see report
in the project folder.
3. Distributed C/S application (Concurrent Reader/ Exclusive
Writer - CREW)
CREW expects following assertions:
1. No 2 writer can access critical region at a time (ie write) - Exclusive
Write.
2. Reader cannot read if writer is writing.
3. Multiple readers can access critical region if they dont violate assertion 2.
- Concurrent Read.
4. Writers cannot write if there are readers reading.
Also, Writers are allowed FIFO access after they are waiting. For futher
details, see report in the project folder.
4. Synchronizing Multithreaded Application
Achieve synchronization between threads for accessing a resource (i.e. entering
critical region) in a multithreaded application. It is also expected that this
doesnot lead to starvation or deadlock.
5. Book Recommendation
Recommendation systems provide users with recommendations of items
based on several similiarity metrics such as past preferences, preferences
of similar user and demographic information. I
have developed a Book Recommendation system based on BookCrossing
dataset which would recommend users a list of books that they might
be interested in. I have studied several state-of-art algorithms for exist-
ing book recommendation systems. I have implemented mixture
models algorithm for recommending books.
6. Using Hive to perform medium-scale data analysis
Pet Projects
1.
GatorMemo
It is a note management system based on spaced learning (Supermemo) algorithm
and mindmaps.
Download version 3.0
code
or
setup (WPF
application)
2. Online Examination using JSP & Servlets
(Similar project - Assignment Assistant)
It uses open source products, therefore is available free of cost. It uses
Apache Tomcat Web Server and MySQL database. Since it uses JSP & Servlet, it is
portable, therefore can be used on Linux as well as Windows. Using Linux
provides University a free testing software.
Responsiblity:
Administrator - managing the account (professor an students)
Professor - Entering/Changing Test Questions, Getting Score Report
Student - Give test and at the end get score report.
Download:
OnlineExamination,
Assignment Assistant
3.
Span Online
ASP.NET/Ajax (and SQL Server) based implementation of Online Examination.
4. Adaptive Traffic Management (ATM)
This project was presented in Redhat Open Source Challenge held at IIT Bombay.
5. Recruit Info
Simple Win32 application that helps HR manager to manage the recruitment
process.
6. Address Book
Simple Win32 application that manages contacts.
7. Word To HTML
Allows you to convert your Word documents to HTML.
8. Automation of Win32 and Web Application (Address book - Win32 application
and Customer Survey - Website)