• Chennai, TN India
Udhayakumar Sundaresan

Udhayakumar Sundaresan

Software Engineer at Entrayn Education Technologies (P) Ltd.

Summary

Software Engineer with close to three years of experience on architecture and developing application which handles gigabytes of user data. Developed advertisement serving engines to participate in real-time bids and implemented machine learning techniques to smartly place the bid's. Worked on implementation of Item Response Theory and calibrated all the parameters involved in it. Gained development and operational expertises on Hadoop, Hbase, Cassandra, elasticsearch, Mahout, Mysql.

Work History

Work History

Software Engineer

Mar 2014 - Present
Entrayn Education Technologies (P) Ltd.
  • Implemented ‘GRE Test prep Math & verbal’ android and iOS app’s backend which serves questions based on learners strength and weakness which is calculated based on his previous attempts.
  • Created recommendation engine which automatically generates a report on users performance and actions user need to take further. 
  • Created the backend for competitive exam practising platform.

Software Engineer

Nov 2012 - Aug 2013
Reducedata ( Acquired by Snapdeal, India )

Charged with design and development of various components of product stack, that serves bid-request at sub-milliseconds spanned across multiple data center.

  • Supply Side Platform (SSP) Integration - Created campaign synchronization and integration of Google AdExchange and MoPub to Reducedata.com RTB platform.
  • Created modules for Ad campaign creation and management using Apache Cassandra database at backend. Designed Apache HBase tables for analyzing the live traffic data of Ad Campaigns.
  • Campaign action recommendation - Designed and created Recommendation module, which monitors live Ad Campaign’s multiple parameters like Daily budget, Total budget and takes/suggest actions based on user-configured conditions. It uses Akka.io to span workers and distribute the workload on HBase cluster.
  • Real-Time Bid Computation Engine - part of Design and Implementation of Real-Time Bid Computation Engine using Real-Time performance data of ad campaign.
  • Programmatic Optimization and Predictive Analysis - Programmatic Optimization and Predictive Analysis with real-time performance data and Machine Learning algorithms using Apache Mahout and vowpal wabbit.
  • High Performance and Low Latency DSP Ad Server - High Performance, High Availability and Low Latency (< 1ms internal processing time) DSP Ad Server using Java 7, JBoss Netty, Java Springs, REST Express, DI, Disruptor Framework, CQengine, Guava cache, Google protobuf and Apache Kafka
  • Audience Targeting - Audience Targeting Feature integrated with various Audience data providers (Semcasting, Bluekai)
  • Real-time Analytics - Real-Time Data Processing and Analytics using Zookeeper quorum, Distributed Kafka, Distributed HBase and Multiple Reporters (distributed data processing system)
  • Cross Platform Retargeting - Leverage the machine learning technology to identify desktop users on mobile devices, Web users within mobile apps automatically irrespective of location. Automatically retarget users across Mobile apps, mobile games and online websites.
  • Scraped campaign from other providers to migrate to reduce data platforms.
  • Actively managed and participated in deployment of Ad-server across different geographic regions.


Skills

Skills

Apache Hadoop

Wrote several pure map-reduce code as well as scalding jobs to handle parquet and sequence file, which stored ad campaign performance data serialized by Avro and protocol buffers.

Wrote several jobs in Pig and analytical queries (join intensive) in HIVE.

Migrated high volumn Hbase tables to S3 store and restored back.

.

Apache HBase

Designed row-keys for more than twenty HBase tables and redesigned them upon requirement changes and migrated data from one cluster to another using pure Map-reduce.Handled several gigabytes of data every day which involves cleaning and sanitization. Planned capacity and handled scaling from three node cluster to twenty node cluster in different stages.

.

Cassandra

Created several ad campaigns data and synced them across multiple data center and CDN's to serve ads at the different geographic region. Integrated campaign sync with DoubleClick Ad Exchange ( Google ).

Java

Designed and Implemented several chained map-reduce jobs to perform large-scale analytics.

Implemented Random Forest and k-Means Clustering using Mahout.

Designed message pipeline using Kafka and HBase.

Scala

Using Akka.io created distributed recommendation engine which works on top of HBase and Cassandra data. Used scalding for several sanitization tasks.

Python

Used for Item response theory implementation and Map-reduce streaming jobs.

Designed web scrappers to crawle web site and store it's content in HBase.

C++

Used for Coding competition in Top Coder and SPOJ.

Created  atomic structure visualizer using graphic.h in High school.

Education

Education

Bachelor of Technology - Information Technology

Sep 2008 - Apr 2012
Prathyusha Institute of technology and Management  

Affiliated to Anna University.

Graduated with First class, with 7.81 CGPA.


Senior school certificate examination

Jun 2006 - Apr 2007
G K Shetty Vivekananda Vidyalaya, Ambattur

Affiliated to CBSE.

Passed with 71%

Memberships

IEEE Student Member 2011 and IEEE Computer Society Member 2011.
Microsoft Developers network (MSDN) 2009 - 2012.

Certifications

Certifications

Introduction to Big Data by University of California, San Diego

Dec 2015 - Present
Coursera

Introduction to Big Data, 1 of 6 courses in Big Data from University of California, San Diego.

Cloudera Certified Developer for Apache Hadoop, License: 100-014-810

Dec 2015 - Present
Cloudera

Individuals who have demonstrated their technical knowledge, skill, and ability to write, maintain, and optimize Apache Hadoop Java MapReduce development projects.

Volunteer Works

Microsoft Student Partners (MSP), Member.
Participated in 3D photography workshop, Pragyan, NITT.
Online Certification GILD| C++ Gold Certified.
Organized blood donation camps inside college, part of college tech festival Xpkye-2011.
Quiz and online coding contest event in-charge, part of college tech festival Xpkye-2011.
Conducted several awareness sessions on organ donations around Tiruvallur district for Farmers.

Research

‘Secured search over encrypted data in Cloud Storage Systems’ - Proceedings of the International Conference on Mathematics in Engineering & Business Management March 9th-10th 2012, Stella Maris College, Chennai, Vol-II, Pg-204-210. ISBN: 978-81-8286-015-5