Microsoft

Engineering lead for Trial engineering team in Customer Service functional group of Dynamics product line and Dynamics Contact center voice platform. 

Architected the Power virtual agent integration into Customer service trial experience. A zero touch experience for trial users, enabling them to use chat and voice channels of Dynamics Omnichannel product. In both these channels, Power Virtual Agent will front the user interaction with an option to transfer the conversation to a human agent. 

Employed data driven decision making to drive user engagement in the trial experience. User research studies on Figma, Telemetry and competitor study were leveraged to inform the decision making process. Philosophy: Less is more for a trial experience. 

Future looking charter for Dynamics Customer service in the journey of transitioning from Sales led revenue growth to a hybrid model of Sales led revenue growth and Product led revenue growth.

Software Engineering Lead

Amazon

For Alexa Authorization, started as a founding engineer in the green field project and grew the platform ground up. I was tech-leading a team of 8 engineers with diverse backgrounds, skillsets, experience levels (15 years experience to Freshers). 
1. Technical lead for Alexa Authorization - Distributed backend systems engineering. Attribute based access control system that serves 200K TPS and provides results with 6ms 90th percentile latency. 


Both Alexa knowledge graph and Duplicate detection projects required using similar technology for their entity linking needs. Data sources varied. 
2. Alexa knowledge graph - Data pipeline engineering 
3. Duplicate Detection (Machine learning, Active Learning workflow, Data pipeline engineering)
SDE in Amazon Catalog Quality team, a part of Search and Discovery organization of Amazon.

Specialities: Spark RDD memory tuning, optimized joins, re-ordering RDD transformations to reduce run-times, feature engineering for Duplicate detection. 

Re-architected and owned a pipeline that runs batch jobs to identify potential duplicates for each entity in the amazon catalog. Improved the cumulative recall to 94% from 80% after 6 consecutive iterations of identifying potential duplicates for members of the catalog. 

The potential duplicates are then processed online using several machine-learned (learning the attributes to look at from the semi-structured entity data) rules that use token similarity measures to determine if the entities should be linked or not.

Stony Brook University

Worked on the evo-grader project at Prof Ross Nehm's labaratory along with Prof. Ross Nehm and Dr. Minsu Ha. For more details, please visit http://evograder.org/

Research Assistant

My Masters thesis is on Semantic parsing and language understanding. Thesis is available here http://www.gthandavam.in/thesis.pdf

MS in Computer Science

Worked on extending a machine-learnable language that was used for expressing rules for solving a binary classification problem in Amazon. 

After extending the grammar in a backward compatible manner, implemented a couple of data-driven greedy optimization techniques to improve the run-time performance of the rule evaluation, attuned to the representative data-set.

SDE - Intern, Search and Discovery

CA Technologies and SUNY Research Foundation

Prototyped the Android Interceptor technology - A platform to enable finer level monitoring of the android mobile apps via CloudMinder/SiteMinder.

Sr Research Aide

The D. E. Shaw Group

Worked in the DailyClose project which handles cash reconciliation.

1. Owned the perl based broker file parsing framework for cash reconciliation
2. J2EE web app for the Cash Reconciliation
 - Added support for Separately managed Accounts (SMA) business workflow
 - Enabled the web-app to interact with services that are behind kerberos wall
 - Exception Report handling module for several workflows in the web-app

Software Engineer

Juniper Networks

Was part of Juniper's SSL VPN team. Worked on the content-rewriter component of the VPN suite provided by Juniper
Worked on: javascript (client side) ;  c++ server side content rewriter engine

Software Engineering Intern

Kalycito

Developed a proof-of-concept module for High availability of Ethernet Powerlink protocol in a team of 2. Done as a part of the "Open Source Systems" coursework.

Passionate hands-on engineering leader with 12 years of full stack engineering experience. Accomplished leader who has delivered products successfully by leading team of engineers, designers, product managers in both Amazon and Microsoft. 

Proven track record of leading engineering teams and working across different functional groups - Program management, Product management, UX Design and Engineering leadership for a product delivery. Partner team communication (peer Engineering leads, UX partner, Program management partner, Product management partner), upward communication (engineering leadership) and downward communication with meaningful context: for team members.

Taplio

Ganesa Thandavam Ponnuraj's Linkedin Analytics

Ganesa Thandavam Ponnuraj

Want to drive more opportunities from LinkedIn?

Taplio

.css-1jgreyo{background-image:linear-gradient(284deg, rgba(39,43,72,1) 0%, rgba(33,47,183,1) 100%);color:transparent;-webkit-background-clip:text;background-clip:text;}Ganesa Thandavam Ponnuraj's Linkedin Analytics

Ganesa Thandavam Ponnuraj

Want to drive more opportunities from LinkedIn?

Ganesa Thandavam Ponnuraj's Linkedin Analytics