SENTHILKUMAR ARUMUGAM

Technology Specialist - BigData

Malvern, PA

senthilka.connect@gmail.com

484-905-1413

http://www.linkedin.com/in/senthilka

Summary

An extensive experience of 6+ years in developing as well as consulting solutions on Java/J2EE, HPC, Big Data, cloud based

applications spanning across technologies and business domains. A Technology Evangelist on various technologies such

as BigData (Hadoop, NoSQL).

Work Experience Highlights

1 Currently Technology Lead in a Service Company - Responsible for Technology Consulting, HPC Initiatives and

Innovation. Area of specialization includes application architectures based on Big Data Technologies (Hadoop

Ecosystem) and NoSQL DataStores (Cassandra, HBase, and DynamoDB).

2 Extensive experience in product development in particular cloud based products.

3 Have mentored the team on various new technologies based on “Big Data”.

4 “Big Data” Enthusiast and Founding member of the tech Community “Chennai Hadoop Users Group”. Have taken a

lot of session on HPC Technologies at corporate as well as academic level. Been a technology speaker at various

academic institutions such as Anna University, VIT University, SSN College of Engineering and many more.

Specialization/Certification

Cloudera Certified Hadoop Developer (CCDH)

Cloudera Certified Specialist in Apache HBase (CCSHB)

Cloudera Certified Administrator for Apache Hadoop (CCAH).

Professional Experience Summary

  · Technology Specialist, High Performance Computing CoE, Cognizant Technology Solutions.(March 2012 -

Present)

Responsible for Technology Assessment, Innovation on Big Data technologies.

Migration of Traditional Applications into HPC based Solutions.

  · Senior Software Developer, OrangeScape Technologies(October 2010 – March 2012)

Lead the team for implementing the automation of runtime deployment of their cloud based product.

  · Consultant, Catalyst EduCare (June 2009 – October 2010)

Involved in the software development as well as consulting for domestic industries. Have mentored the research

scholars for publishing their research papers

  · Member Technical Staff, HCL Technologies (October 2007 – June 2009)

Involved in the development of the HCL VoIP Product

Education

  · Bachelor in Engineering, Electronics and Communication, 2007, India, First Class with Distinction

Area

Operating Systems

Cloud Platforms

BigData Technologies

Hadoop EcoSystem

Hadoop Distributions

Linux – Centos, RHEL5, Windows

AWS, Google App Engine, OrangeScape

Hadoop Ecosystem, Storm, ElasticSearch, Solr, Kafka

HDFS, MapReduce, HBase, Hive, Pig, Oozie, zookeeper, Sqoop, Impala, Phoenix, Omid,

Flume, HCatalog

Cloudera, Hortonworks, MapR

Skillset

Languages

NoSQL DataStores

Amazon Web Services

Web Technology

Methodology

IDEs, Frameworks, Tools

Servers

Caching Technologies

Protocols

PROJECT DETAILS

Organization: Cognizant Technology Solutions

Project #1

Title

Client

Duration

BigDataTeam Size

Hadoop Ecosystem

Java/J2EE, Python, C/C++, C# .NET, HTML, DHTML and JavaScript

HBase, Cassandra, DynamoDB, BigTable

EC2, S3, SimpleDB, RDS, DynamoDB, EMR, VPC

Spring MVC, Pylons, Restful services

Agile Scrum

Apache Jena, JMS, SpringBatch, Maven, Ant, Ecplise IDE , Selenium, JMeter, Ethereal, GCC,

Boto Library, MS Office and Adobe Apps

Apache Tomcat, JBoss, Tornado, GlassFish

MemCache, IBM eXtremeScale

TCP, UDP, SIP

Siemens’ Care Management System (Product)

Siemens

OnGoing Projec from June 2013

20

Programming Languages

Server

Other Frameworks

Operating Systems

Hadoop Distributions

Hadoop, HBase(M7) , Hive,Sqoop,Zookeeper, Hive , Impala, Phoenix

Java, Spring MVC

Apache Tomcat

Apache SolR(LucidWorks), TIBCO EMS , Apache Jena

Linux, Windows

Cloudera(CDH), MapR M7

Project Objective:

  · Building data store to store various types of Health Care Events

  · Provide retrieval of the data across the system including UI

  · Evaluate the performance and scalability of the solution

  · Build real time search over the data stored in DataStore.

Project Description:

Modules under this project are:

  · Design and implement the Persistence layer of the events in DataStore for batch and online retrieval

of the data.

Roles and Responsibilities:

  · Analysis of bleeding edge technologies , tools for better architecture.

  · Design and Implementation of Deployment Architecture

  · Base layer Coding and mentoring the team on development

  · Performance Tuning

  · Analyse and implement the ontology tree using Apache Jena for executing business CEPs.

  · Retrieval Layer for various system components such as UI(app Server) and Tibco Components

  · Evaluate the other components such as Impala, OMID, Phoenix for the system.

  · Identifying and resolving scalability bottlenecks

  · Responsible for choosing the right vendors for Big Data technologies.

Project #2

Title

Client

Duration

Team Size

Hadoop Ecosystem

Operating Systems

GDM Data Quality Report Benchmarking

Dun & BradStreet, Inc

3 months

5

Programming Languages

Server

Tools

Hadoop, HBase, Hive,Sqoop,Zookeeper

Java

Apache Tomcat

Apache Jmeter

Linux, Windows

Project Objective:

  · Building and benchmarking of data quality report generation process from GDM datastore using

Hadoop Ecosystem Technologies.

Project Description:

Modules under this project are:

  · Evaluate the performance and scalability of the solution

  · Build a RESTful services for online retrieval of data and perform load testing on the server

  · Generating IDBar and Bankruptcy reports along with Batch Fullfillment and online retrieval of GSRL8

data using given set of DUNS.

  · Implement the report generation process for sample reports: IDBar and Bankruptcy Events using

Hadoop, Hbase and Hive tech stack.

  · Designing and implementing an approach for batch and online retrieval of GSRL8 data in CSV format

for the given set of DUNS numbers using Hadoop stack and REST services.

  · Evaluate the scalability of the solution

  · Key technical aspects that are validated :

o Approach of report generation with GSRL8 data stored as individual MPFs and as entire XML.

o Tuning of Hive UDAFs for improved computing task throughput and performance.

o On storing GSRL8 data as multiple MPF XML fragments, DUNS number and MPF name will be

used for updating specific MPFs only.

Roles and Responsibilities:

  · Requirement Analysis.

  · Coding

  · Environment Setup (Hadoop, HBase and Hive Clustering)

  · Performance Tuning

  · Identifying and resolving scalability bottlenecks

  · Scaling of process from 1 hardware node to 7 nodes

Project #3

Title

Client

Duration

Team Size

Hadoop EcoSystem

Lilly APLD analytics

Lily SnowBall

2 months

2

Programming Languages

Operating Systems

Hadoop, Hive,oozie

Java, Python

Linux, Windows

o Data replication between nodes and consistency

Project Objective:

  · To evaluate analytics on Hadoop EcoSystem for Lilly Snowball and validate functionality of APLD data

  · Evaluate the performance and scalability for large volume of data.

  · To recommend hardware configuration for performing analytics in the client environment.

Project Description:

Modules under this project are:

  · Implement the report generation process for sample reports: Patient Continuity, New to the Market, and

4 more batch jobs using Hadoop and Hive tech stack.

  · Implement the margin based CPO allocation process, using open source stack (Hadoop, Cassandra )

  · Key technical aspects that are validated :

o Workload distribution and parallel processing

o Compute task throughput and performance

Roles and Responsibilities:

  · Requirement Analysis.

  · Coding and Data Generation

  · Environment Setup (Hive and Hadoop Clustering)

  · Performance Tuning

Project #4

Title

Duration

Client

Team Size

Hadoop Ecosystem

Amazon Services

  · Identifying and resolving scalability bottlenecks

  · Scaling of process from 1 hardware node to 8 nodes

Cognizant’s Intelligent Archive (Product)

4 month

Dun & BradStreet, Inc

4

Programming Languages Java

Operating Systems

Hadoop, Hive, Oozie, Cassandra

DynamoDB, SimpleDB, EC2, S3

Project Objective:

  · To Archive large volumes of structured data and provide low latency of the data

  · Migrate given huge volume of data to HDFS/ Amazon S3 and introduce indexing on it.

  · Create views and store the corresponding metadata in Cassandra / DynamoDB.

  · Evaluate the performance and scalability for large volume of data.

Linux, Windows

Project Description:

Modules under this project are:

  · Data Migration: Move the data to either HDFS or Amazon S3 using Amazon API

  · Indexing: Create indexing on each file like the start and end bytes of each record using MapReduce

  · View: Create the corresponding views and fix the search parameters

  · Metadata Generation: Extract the respective metadata for the views and store it in Cassandra/

DynamoDB.

  · Scalability: How to scale up the system for large volume of data for ingestion using MR as well as

Storage.

Roles and Responsibilities:

  · Requirement Analysis and Coding

  · Scalability as well as Performance.

  · Retrieval: View is generated by retrieving the stored data and the retrieval is generalized in such a way

that it need not be rewritten for new views.

Project #5

Title

Duration

Team Size

Hadoop Ecosystem

Operating Systems

Cognizant’s ExtremeBatch(Product)

4 month

4

Programming Languages

Others

Hadoop, HBase, HDFS, MapReduce, Pig, Hive, Oozie, Cassandra

Java/J2EE, SpringMVC

Quartz, IBM

Project Objective:

  · To develop a generic framework to solve most of the problems of BigData

Linux, Windows

Project Description:

Modules under this project are:

  · Data Ingestion : To ingest data with varying varieties at higher consumption rate

  · Stream Processing: Create indexing as well as applying rules during persistence of the data

  · Polyglot Persistence: Providing the flexibity to choose the storage for each app layer as well as client

centric storage

  · Scalability: How to scale up the system for large volume of data for ingestion using MR as well as

Storage.

Roles and Responsibilities:

  · Technology Assessment of all the technologies

  · Requirement Analysis and Coding

  · Scalability as well as Performance.

  · Documentation of the Approaches/technologies for future consulting needs.

Organization: OrangeScape Technologies

Project #6

Title

Duration

Team Size

Amazon Services

Programming Languages

Servers

PAAS

Operating Systems

S3, EC2, RDS

  · Central Management: To create base portal for central management of the both batch as well as real

time processing

OrangeScape Studio(Product)

17 months

11

Python, Java, Jquery, C#

Tornado, JBoss, GlassFish

Google App Engine

Windows, Linux

This Project involves the engineering enhancement of PAAS product OrangeScape Studio. The product is a

visual PAAS over Google App Engine and applications developed can be deployed across any platform or

clouds.

Roles and Responsibilities:

  · Implemented core middleware components of the Product such as authentication.

  · Moved the complete build process of the product to EC2 and S3.

  · Implemented the click once deployment of the application from the product using S3 and Tornado

Server.

  · Designed and implemented the selenium based automated as well as manual testing of the product on

EC2 as on demand computing.

  · Played a vital role for implementation of Agile Scrum cycle.

  · Have also bundled a paid AMI for all application built using OrangeScape.

  · Enhanced performance monitoring of the cloud application developed using OrangeScape Studio.

  · Have provided constant support for the onsite team.

  · Have developed UI components using HTML, JavaScript, CSS, JSON, and JST.

  · Implementing the debugging and logging feature for the application developed in OrangeScape Studio

across the platforms.

Organization: HCL Technologies

Project #7

Title

Duration

Team Size

Programming Languages

Protocols Used

Operating Systems

C

TICKET MANAGEMENT SYSTEM

6 months

1

SIP,TCP,UDP

Windows, Linux

This project involves notification of the tickets received in the HCL Sip Server from the WFFM [wireless field

force management] to the sip user. The server receives the ticket using unconnected UDP socket and SIP/SIPS

unsolicited NOTIFY is built. Then the built message is sent to the user if he is online. The tickets will be stored in

the sip server in memory if the user is offline. It will be sent when he logs in.

Roles and Responsibilities:

  · Programing

  · Testing and Bug Fixing

Project #8

Title

Duration

Team Size

Programming Languages

Protocols Used

Operating Systems

SERVER RELEASE TESTING

3 months

8

C

SIP,RTP,SDP,TCP,UDP

Windows, Linux

This project involves the testing of the HCL Sip server with respect to the Test plan. Server was tested for its

RFC3261 compliance and also other RFCs .All the services such as presence, registration and other services in

the Sip server stack was tested extensively.

Roles and Responsibilities:

  · Programing

  · Testing and Bug Fixing

Project #9

Title

Duration

Team Size

Programming Languages

Protocols Used

SERVER REGISTRATION EVENT PACKAGE

3 months

3

C

SIP,RTP,SDP,TCP,UDP

Operating Systems

Windows, Linux

This project involves the Reg event package as a service over Sip Stack. The Service handles the Subscribe

message event ‚reg‛ as per RFC3680. With this, its possible to share the bindings stored in Registrar with other

application or machine. It must sit with Registrar.

Roles and Responsibilities:

  · Programing

  · Testing and Bug Fixing

Organization: Catalyst Educare

Project #10

Title

Duration

Team Size

Tools

Operating Systems

Research On ‘MANETS’

6 months

1

NS2, GlomoSim

Windows, Linux

central management. In MANET, every node acts both as a host (generates user and application traffic) and a

router (carries out network control and routing protocols).To Propose a new protocol for routing of information

on MANETS.

Mobile ad hoc networks (MANET) consist entirely of mobile nodes that communicate on-the-move without

Roles and Responsibilities:

  · have analysed and completely proposed a new protocol based on cluster based routing as well as

location information of the nodes.

  · Simulated the same protocol for the parameters such as end to end delay, control overhead, packet

delivery ratio.

Project #11

Title

Client

Duration

Team Size

Programming Languages

DataBase

Operating Systems

NCVT Management Application

EduGuru

6 Months

4

C#, ASP.NET

Windows 2003 Server

SQL Server 2005

An Intranet web based application for MES courses (dget.nic.in/mes) conducted by EduGuru

Technologies, Chennai. All the students details are entered by all the authorized persons and reports are

generated as per the test results given by assessment bodies.

Roles and Responsibilities:

  · System Architect

  · Client Interfacing

  · Programing and Delivery

  · The research papers are selected in various journals such as ETRI, theSciPub and various conferences.