• Integrated Cisco HyperFlex snapshots into Cohesity’s VMware virtual machine backup workflows to leverage more efficient snapshots.• Massively parallelized Ops (tasks) and implemented fine-grained locking to make Pulse(component to which other Cohesity components report to)scalable. Added crash injection testing to Pulse.• Worked on migrating the backing storage of Stats component from in-memory to Scribe database(key-value store) to alleviate the problem of high memory consumption. Implemented the new Ops (tasks) to be lock-free for usage on a distributed cluster.Used C++, Protobuf, gRPC

• Teaching Assistant under Prof. Scott Smolka for CSE305 (Principles of Database Systems) Fall 2017 for senior undergraduate students.• Conduct office hours every week to help students with database projects and assignments.• Grades home works and exams

Member of a team responsible for developing and maintaining RMAN(Recovery Manager) and ZDLRA(Zero Data Loss Recovery Appliance), the backup and recovery products of ORACLE.• Built 'restore range' functionality for ZDLRA to calculate the recovery window of a database that determines the time in the past to which the database can be restored.• Added a feature to Oracle 'Block Change Tracking' by making it work through 'resetlogs'(an incomplete database recovery) to create faster backups by reducing I/Os.• Provided 'RMAN SDK' for Oracle 12c. This SDK is the minimal test infrastructure provided by RMAN to SBT library vendors to test their tape backup libraries.• Implemented RMAN 'client throttling' for ZDLRA which limits RMAN connections to ZDLRA based on various parameters depending on ZDLRA's capability.• RMAN creates backup pieces which have their own format. Improvised the existing 'piece dumper' to read the magic blocks of a backup piece and write these values to output in an understandable format.Used C, SQL, PL-SQL languages and SQL*Plus tool during development.

• Built Geocoding and Reverse Geocoding scrapers to convert a set of addresses into latitude and longitude co-ordinates and vice-versa• Created a Scrape Validation tool which analyses a scrape, compares it with previous scrape and provides snapshots at a level of city, state etc. The tool automates various aspects involved and sends a report to the entire team through email• Used existing Local Scraper Parsers to extract PBA(Phone Book Answer) results from scrapes of various non-local segments and created local query sets from non-local sets