Jan 2021 – Nov 2022 · 1 yr 10 mos · Bengaluru, Karnataka, India

• Was site lead for AWS EMR team in Bangalore and grew the center from scratch to 45+ over a span of ~2 years. The EMR team in BLR owned independent charters for Presto, Hive, Hadoop, Iceberg and Interactive Workloads/Notebooks for EMR on EKS and EMR Serverless. Was responsible for defining the charter, roadmap, deliverables, growth and operational excellence.

Oct 2017 – Jan 2021 · 3 yrs 3 mos · Bengaluru Area, India

• Qubole offers big data as a service on the cloud. Primarily responsible for leading and growing 5 backend engine teams - Presto, Quantum (Serverless SaaS offering around Presto), ODBC/JDBC Drivers, Partner Integration and Backend engine performance.
• Established and grew these teams from 5 to 22 people across geographies while scaling up the overall product usage 10X from 182K compute units to 1.93M compute units and from a single AWS environment to 9 environments across AWS, Azure and GCP.
• During this time, Qubole Presto team did great work both in open source (Dynamic filtering and Hive-ACID connector) as well as in proprietary Qubole stack (a rewrite of the Presto admission control and scheduling system to be able to run more concurrent/bursty Presto workloads).
• Was responsible for initiating and developing our relationship with Presto founders (Martin Traverso, David Philips and Dain Sundstorm) which eventually helped us organize India’s first ever Presto Summit in 09/2019 with Presto founders.
• Have also been responsible for overall technical evangelism around Qubole Presto via presto@qubole handles across Twitter, LinkedIn and Facebook.
• Was intimately involved in the design, development, testing, launch and operation of our Presto based serverless offering - Quantum (Qubole’s version of AWS Athena).
• Led the Qubole ODBC/JDBC drivers team to do multiple redesign iterations of Qubole drivers to release successive new generations of drivers with significantly improved performance and stability.

Jun 2016 – Sep 2017 · 1 yr 3 mos · Bengaluru Area, India

• Instart Logic was the world’s first endpoint-aware application delivery solution (similar to a content delivery network (CDN)) that makes websites and applications fast, secure, and easy to operate. Instart Logic was funded by Andreessen Horowitz, Greylock, Kleiner Perkins, Stanford, Telstra, and several other notable Silicon Valley investors. I led the agile product development agenda (Jira, Scrum, Gerrit code reviews, Jenins release jobs) for the core CDN infrastructure. Specifically I was involved in the following:
• Design and development of key CDN infrastructure pieces that are responsible for fetching webpages from various origin webservers. Designed protobufs and gRPC infrastructure and added code (C++, Java and Python) for improving logging and monitoring capabilities, as well as for parsing of customer configuration inside Nginx proxy servers. Developed a domain specific language (DSL) for customer configuration and generated parsers for it in various languages using Antlr.
• Design and development of a text detection filter for images – I evaluated and compared text detection in images using OpenCV as well as through deep learning methods such as convolution neural networks. Images with text represented specific challenges to InstartLogic's image processing pipeline and we needed to detect such images and handle them differently.

Jun 2003 – Jun 2016 · 13 yrs

• I was a Senior Technical Staff Member (STSM) and Research Manager at IBM Research.
• The title of STSM is bestowed by the IBM Corporation upon key IBM employees who have made significant and vital contributions to the IBM team, beyond those expected of senior engineers, programmers, and scientists.
• Bootstrapped the SDN effort at IBM Research India and grew the team to 10. Responsibilities included setting up the technical agenda for cloud and networking teams, after discussions with various stakeholders, conceptualising research project proposals and securing funding for them, recruiting top quality Research Staff Members and Software Engineers, conducting regular feedback and 1:1 meetings with team members, participating in technical, architectural and design discussions, conducting code reviews, conducting performance appraisals, and update various executives in different parts of IBM on a regular basis.
• People management responsibilities of a team comprising of PhDs from Ivy League colleges in the US, experienced Software Engineers from top engineering colleges in India and the US, fresh-out-of-college grads from top IITs in India.
• Provided technical guidance to the following efforts:
• Docker container cloud – Docker Image Analysis, Docker Storage Migration, Docker High Availability and Single Point-of-Failure, monitoring and problem determination
• OpenStack Troubleshooting and Problem Determination
• Software Defined Networking use cases (multicast, load balancing, VM mobility, bandwidth reservation) and challenges (Security of SDN Networks, Security of SDN Controllers, Reduction of switch state).
• In the past, I worked on enterprise Java performance, impact of OS Jitter on tightly coupled parallel programs in large clusters, system management and health monitoring in WebSphere eXtended Deployment (XD) and on decentralized orchestration of composite web services.
• Published ~40 papers in well known conferences/journals and filed 27 patents.

Feb 2002 – May 2003 · 1 yr 3 mos

• Job responsibilities included design, implementation and reengineering of their portfolio analysis applications. Work mainly involved redesigning, reengineering and porting the backend C++ server process from Solaris to Linux. Was responsible for handling various issues arising due to differences in the threading model, signal handling, and byte ordering on Linux. Prior to joining the portfolio analysis group, attended the 4 month long intensive Distributed Systems training program covering Unix, Advanced C++, OOAD, Perl, SOAP, XML and basic financial concepts.

Sep 2001 – Jan 2002 · 4 mos · New Jersey

• Worked on the GlobalDecisionMaker Project for Dun & Bradstreet, Inc. This application had JRun as the servlet/jsp engine and WebMethods as the application server. Job responsibilities included working on JSPs/Servlets for adding new features and enhancements.

Sep 1999 – Jul 2001 · 1 yr 10 mos

• Worked part time on various research projects, including:
• a) Design, implementation and experimental evaluation of a prototype middleware substrate that enabled interoperability among multiple, geographically distributed instances of computational, interaction and collaboration servers through CORBA.
• b) Dynamic content (CGI) request scheduling algorithm and cooperative caching of dynamic web pages on a cluster based distributed web server.
• c) Binary Rewriting of Java classes at runtime to make an untrusted class remotely accessible through RMI.