Resume
Dinesh Agarwal
Dinesh Agarwal
817,Twin Oaks Dr,
Apt #1,
Decatur – 30030,
Georgia ‐ USA
Tel: 1‐678‐677‐2549
E‐mail: dagarwal2[at]gsu[dot]edu
________________________________________________________________________________________________
Education
*PhD in Computer Science, Georgia State University, 2008 to present
GPA: 3.73/4.0
Research Interests: My current research centers around creating efficient data structure
for multicore, manycore and GPU architectures.
Brief Synopsis of research:
There are many parallel applications such as Discrete Event Simulations, Branch-and-
Bound, Event Queue Management, task scheduling, etc. that need priority based data
structures. Existing work has studied many such data structures on traditional parallel
architectures. These approaches do not leverage the power of emerging architectures. My
focus is on utilizing recent architectures, specifically multicores, manycores and Graphics
Processing Units (GPUs), to parallelize such data structures. Goal of this research is to
create a framework to facilitate seamless integration of priority based data structures with
any parallel application. On a big picture we are trying to contribute, to the parallel
computing community, some tools that make it easier for parallel computing to be
adapted at all education levels starting as early as at high school.
*MS (IT), Dhirubhai Ambani Institute of Information and Communication
Technology, 2005 - 2007
GPA: 3.67/4.0
*BIT, Jiwaji University, 2001-2004
GPA: 78.48/100
Employment
*December 2009 – Present, Research Assistant, Dept. of Computer Science, Georgia
State University, USA.
*August 2008 - July 2009, Research Assistant, Dept. of Computer Science, Georgia
State University, USA.
*August 2009 – December 2009, Instructor for course “Introduction to Computer
Programming for Non‐Majors” (Python and XML). Dept. of Computer Science,
Georgia State University, USA.
*July 2007 – July 2008, Programmer Analyst, Mindtree Ltd., India.
Skills
General skills in data analysis and research project management. Specific interests
and expertise in:
*Computing Skills
Programming and Scripting Languages: C, JAVA, Python, Prolog, Perl, DRC,
Datalog, Visual Basic, SQL, PL/SQL, JavaScript, HTML, XML, XSLT, PHP, JSP,
VBScript
Operating Systems: Windows, Linux, Mac-OSX
Databases: Oracle, MySql, DB2, CloudScape, MS-Access
Tools: MicroStrategy, Rational Rose Suite, Crystal Reports, Informatica, Business
Objects, IBM WebSphere Studio, Macromedia Dreamweaver, Adobe Photoshop
*Communication Skills:
Taught “Introduction to Computer Programming for Non‐Majors” during Fall 2009
semester.
Took “Teaching Pedagogy” course during Fall 2008 semester.
Won several prizes for my acting acumen.
Papers Presented
*A Robust Service oriented Architecture based solution for GIS Interoperability:
Presented at Map-Asia 2007, an International Conference held at Kuala Lumpur,
Malaysia, this paper presented a comprehensive model to gather spatial data from GML
files and present it over internet in the form of SVG(Scalable Vector Graphics) files. The
paper discusses a robust solution, which used WMS and WFS web services, using which
a user can access text-based spatial data over a web browser in the form of interactive
maps. The transformation from GML to SVG was achieved using XSLT style sheets and
ecmascript.
*Grid Computing - An overview: This paper was presented at National Conference on
Information & Communication Technology (NCICT – 2005). The paper introduced
possibilities of sharing the networked resources outside a virtual organization (VO). In
addition to that, it covered the usage of Grid computing to meet the challenges in data
mining from distributed databases.
*Reloaded Grid – Building Competitiveness in the interconnected world: This paper
was presented at a national level conference named AIMS Western Region Conference
during 2005. The paper was co-authored by a second year MBA student. It discussed
about the efficacy of Grid Computing in providing optimum usage of available resources
for any organization from managerial perspective.
*MindTree Bus Route Management System using Web Based Geographic
Information system: This paper proposes a solution to solve the current problems with
the system being used by MindTree Ltd. This paper won the first prize in annual
technical festival called Osmosis at MindTree Ltd.
Industry Projects
*BlackHawk Network Enterprise Reporting
Client: BlackHawk Network, USA
Period: July-07 to May-08
Description
Blackhawk, a subsidiary of Safeway Inc wants to build an enterprise DW that will cater
to all kinds of reporting needs of the user community, both internal and external. The
project will provide BlackHawk with an enterprise reporting platform for their gift card
business and other subject areas like Finance and Supply Chain as well as new product
lines like Fast forward and Health cards
Role
Team Member involved in enhancement and Reporting. Team Member involved in
support and Maintenance of existing reports.
Responsibilities
• Conversion of specification to reports
• Unit / System testing
• Handling maintenance and enhancement requests
• Fixing Bugs raised by users
• Deployment of the development to QA and production systems
Tools
• Aqua Data Studio
• Data Stage
• MicroStrategy
*Spatial Data Management using XML
Client: Bhaskaracharya Institute for Space Applications and GeoInformatics, India
Period: January-07 to May-2007
Description
BISAG collects satellite images of earth in the format of shape files, which should be
available to a user with a web browser in the form of interactive maps. To make the maps
interactive the data should be displayed in SVG (Scalable Vector Graphics) form. The
project will provide BISAG with a comprehensive tool, which can collect data from
shape files, organize it in GML files, convert it to SVG files and send it over the Internet.
Role
Architect for XML schema and related XSLT stylesheets.
Responsibilities
• Processing shape files and to convert them to GMLfiles.
• Creating and maintaining Web Services that transform GML to SVG using XSLT.
• Writing Scripts for user interaction
Tools and technologies
• Microsoft C#
• XML Writer
• XSLT
• SVG
• Macromedia Dreamweaver
• ECMA Script
Academic Projects
*Parallel Priority Queues
This is a research project that I am working on with my Advisor Dr. Sushil Prasad. We
are trying to implement Parallel Priority Queues on multicore architecture. A Parallel
priority queue is a parallel heap with node-capacity r > 0 is a complete binary tree such
that (i) its each node contains r items (except its last node which may contain fewer
items), and (ii) at each node all its items have values less than or equal to the items at its children. When the second constraint holds at a node, the node is said to satisfy the
parallel heap property, which is an extension of the ‘heap property’ for conventional
heaps. The parallel heap property, when satisfied at all the nodes, ensures that the root
node of size r of a parallel heap contains the smallest r items (the r highest priority items). A parallel heap with node-capacity r employs r processors. I am trying to modify the existing algorithms to adapt Parallel Priority Queues to harness the power of multicore
processors. Subsequently, we are going to implement Parallel Priority Queues on GPUs
to leverage their massive parallelism.
*DRC query interpreter
The objective of this project was to take a DRC query and check it for any errors. The
modules in this project were to check if the given formula is well formed or not, if the
query is OR safe or not, getting all the sub-conjuncts in the query, to check if the query is safe or not and finally to check if the query has a safe schema or not. This project was
implemented in PROLOG.
*Parallel Adaptive Page Rank
Page rank is a link analysis algorithm used by Google internet search engine that assigns
a numerical weighting to each element of a hyperlinked set of documents, such as World
Wide Web, with the purpose of measuring its relative importance within the set. In this
project we improved on regular page rank by making it adaptive. A page’s rank is only
calculated if it has not converged till that iteration/pass. There were two versions of this algorithm that were implemented. One was sequential algorithm and other was parallel
algorithm that can run on any arbitrary number of processors. The parallel algorithm
itself had two versions, in which the advanced version created threads on each processor
thereby reducing the inter-process communication. The algorithm was implemented using
C with MPI as message passing framework.
*Content Management System
Content Management System, as the name suggests provides a solution to publish
information in an elegant manner over Internet. The project has the functionality to take
information directly entered into the underlying database, or reading it from flat text files called metadata files. Once the data is collected it is stored in a metadata repository and based on this a user can create any number of web pages for different content conforming to same schema. The project was capable of creating HTML forms, which can create specialized HTML elements like text boxes allowing only numeric or alphabetic data, check boxes and radio buttons. Also a user can create forms that can accept a file as
input. The authentication module was linked with the university web-mail and a student’s
official email id and password can be used to access the project web pages.
*Smart Grade Manager (SGM)
SGM was created under the supervision of Professor Dr. R.B. Lenin. He provided us with
the requirements of a faculty member to Asses a student’s performance. SGM enables a
faculty member to maintain academic details related to various courses and students
enrolled in those courses in an efficient way. There are mainly three types of assessments
in a course - quizzes held in class, internal or external examinations and
laboratory/practical examinations. Each of these examinations can contribute some
percentage to final marks (e.g. each quiz being of weight 5%), or they may represent
certain percentage when aggregated (e.g. all quizzes contribute towards 15% of final
marks). The scores can be imported from an MS-Excel spreadsheet or can be manually
inserted. SGM has the ability to project results in the form of Graphs. It can also display
the minimum, maximum, average and standard deviation of the marks among students.
SGM was created using Visual Basic 6.0 as front end and MS-Access 2003 as back-end.
The reporting tool used was Crystal Reports 8.5.
*ShabdaPurna
ShabdaPurna was created under guidance of Professor Abhinay Pandya at DA-IICT.
ShabdaPurna, which means, “Complete the word” in Sanskrit, is an online multi user Scrabble
gaming portal. A user needs to register to play; once the user logs-in he can select an
opponent from the list of online players. When both parties agree a new game starts, the
winner is the person who scores 100 points early. The user can create a word in any of
the eight directions in a 10 X 10 matrix. The word claimed by user was check against a
Standard English dictionary available on server. The ShabdaPurna is developed using
PHP with MySql as database. Apache server was used to host the game website.
Personal Achievements
• Topped University in BIT (Bachelor of Information Technology) 3rd Semester.
• Secured 3rd rank in MS-IT (Master of Science – Information Technology).
• Topped University in MS-IT 1st semester with a SPI (Semester Percentage Index) of
4.0/4.0.
• Two times First Prize winner at state level Proscenium competition
• Second prize in street play during annual techno cultural festival at DA-IICT
• Awarded ‘Best Actor’ during Synapse 2006 (annual cultural festival) at DA-IICT.
• Served as Radio Jockey for ‘Radio – DA-IICT’
• Active member of DA-IICT cultural committee, as I have lived in many different states
and cultures, I helped institute in establishing cultural harmony among students by
organizing cultural programs.
• Successfully lead the theatre group “Khoj” at DA-IICT, which today is a well-known
identity at all national level competitions.