Sameer Kedia Email & Phone Number

London, England, United Kingdom

London, England, United Kingdom

Bengaluru Area, India

Cluster Management Team- Currently leading the team's effort to upgrade the version of python used in Qubole’s code base from 2.7 to 3.8.- Added support for migrating shuffle data in Qubole’s GCP offering by restarting preempted nodes. This ensures higher availability of data causing less failures because of instance preemption and better execution times for big data jobs.- Guided other team members on important projects to enhance security in QDS - support for the newer Amazon Linux 2 image to launch clusters and replacement of use of SSH by APIs to access clusters.- Added support for GCP in QDS. Created cluster orchestration layer, automated creation and publishing of machine images with Qubole packages, explored various possible use cases of GCP features and ramped up other engineers on GCP related concepts and technologies.- Added support for newly introduced instance types which provide higher computing power enabling the customers to run their big data jobs faster.- Added support for selection of the cheapest subnet to launch clusters for different types of configurations. This effort helped bring down the customers’ cloud costs.- Worked on a refactoring project to have a common orchestration layer for all the supported clouds - AWS, Azure, OCI and GCP. This allowed easier and seamless addition for support of new clouds and less developer time spent on repetitive code bases.Middleware Team- Owned the project to upgrade the tunnel infrastructure in QDS to enhance security by adding support for user configurable SSH user and port. This required end-to-end development including UI, collecting data on potentially impacted customers, communicating the migration steps and educating others.- Worked as a sole member of the data import and export team. Added a number of features and improvements to data import and export in QDS - efficient and faster data transfer to and from AWS Redshift, MongoDB support and support for BLOB data type in SQOOP.

Bengaluru Area, India

- Upgraded the version of open source tool SQOOP used in Qubole Data Services from 1.4.2 to 1.4.6. SQOOP is used to transfer data between Hadoop and structured datastores.- Upgraded SQOOP to use the latest version of Hadoop and added support for running data import and export jobs on clusters in the customer's account instead of a shared cluster to enhance security and efficiency.

Mumbai Area, India

- Developed the admin module for the Bootstrap web portal, to allow admins to grant/revoke access to reports and dashboards for employees.