Oct 2021 – Oct 2023 · 2 yrs

• Designed and led metadata synchronization framework between Rubrik's SaaS product and on-prem product.
• Improved sync latencies of critical user workflows by up to 180x, and reduced redundant work done by the framework by up to 9x.
• I lead projects in the job framework team, and introduced APIs to improve job dispatch throughput by up to 360x.
• Designed the next generation framework to sync metadata between the on-prem and SaaS product to reduce sync latencies from 10s of minutes to seconds.
• I lead the design review form discussions across the platform team.

Apr 2019 – Oct 2021 · 2 yrs 6 mos

• Primarily led projects in Rubrik's distributed database and distributed filesystem.
• Blog on migrating Rubrik's distributed database from Cassandra to CockroachDB
• 1. https://www.rubrik.com/blog/architecture/22/3/choosing-the-right-metadata-store
• 2. https://www.rubrik.com/blog/architecture/22/4/choosing-the-right-metadata-store-part-3
• 3. https://www.rubrik.com/blog/architecture/22/3/choosing-the-right-metadata-store-part-2
• Lead cross component projects to develop a new logical file format for improved restore performance. This project reduced the time an ingested snapshot is available for restore by 61x.
• Introduced “range filters” to a file format to improve read performance by up to 50%.
• Introduced novel physical file encoding format which can delete unreachable data in constant time without requiring to rewrite files.
• Used cutting edge tracing tools to improve performance of the distributed filesystem. Improved performance of writes by up to 7x by analyzing these traces and improved reads performance by up to 50%.
• Also improved performance of a snapshot download from a cloud archival location by 10x.
• Worked on a cross component project to support rollback after an upgrade which also involved adding a novel approach for snapshotting and restoring our distributed filesystem.

Aug 2017 – Mar 2019 · 1 yr 7 mos

• Founding member and a technical lead of Callisto India. This team works on Rubrik's distributed database. We made open source contributions to CockroachDB to increase its resilience and fault tolerance.
• Worked on validating CockroachDB as a candidate for a distributed database and implemented generic load generation frameworks https://github.com/scaledata/rksql.
• Designed and lead a team to build a distributed monotonic clock: https://github.com/rubrikinc/kronos to integrate in CockroachDB to protect it against clock skew and clock jumps.
• Designed and implemented copysets in CockroachDB to significantly improve tolerance to multi node failures.

Jul 2016 – Aug 2017 · 1 yr 1 mo · Bangalore

• Worked in developing the platform for the e-commerce advertising team. Primarily worked on building novel algorithms for the bidding and product recommendation engines.
• Designed and deployed an online Image Processing service to dynamically resize and pad images
• Created and implemented bidding algorithms to optimize for clicks and transactions
• Integrated the bidders with both internal and external Ad-Exchanges
• Conceptualized and made live a new novel generalizable algorithm for click prediction
• Built tracing tools to analyze platform performance and increase scalability
• Created a budget management system which avoids over / under spending
• Handled major data migrations across stores and data centers with no downtime
• Prototyped catalogue ingestion using Spark which increased our throughput by more than 50 fold
• Designed and implemented a product recommendation system based on user cohort

Aug 2015 – Jul 2016 · 11 mos · Bangalore

• Created alerting and monitoring systems to monitor health of production traffic in all data centers
• Designed novel Machine Learning Models for user product recommendation
• Set up a cross data center replicated catalogue storage database using Elasticsearch
• Dockerized Inmobi’s Remarketing Ad-Serving applications

Jun 2013 – Jul 2015 · 2 yrs 1 mo · Bangalore

• Designed an algorithm to find and an resolve conflicting constraints in an optimization problem
• Created a tool to monitor realtime statistics for a market index using a client-server architecture
• Designed an infrastructure to add custom data integrity rules for raw and processed data
• Enhanced a tool used for research by adding support for new kinds of datasets
• Implemented signals used in quantitive models for investing

May 2012 – Jul 2012 · 2 mos · Redmond

• Enhanced Transliteration tool for Hindi. Reduced error rates by half and improved the speed 7 fold
• Trained Transliterators for Tamil and Urdu
• Wrote a tool to filter out possibly parallel dynamic websites from a list of pairs of websites and
• wrote targeted crawlers for parallel websites
• Used deep learning techniques in Sentence Alignment

May 2011 – Jul 2011 · 2 mos · Bengaluru Area, India

• Implemented an algorithm to compress a large number of images with minimal quality loss by looking at frequencies present in images
• Developed an algorithm to display images of different sizes in Search with least whitespace between them
• Developed a Dynamic Programming Algorithm to intelligently crop images
• Worked on extracting information from metadata of images using Hadoop

Aug 2009 – May 2013 · 3 yrs 9 mos · Chennai

• Cumulative Grade Point Average (CGPA): 9.75
• Awarded the President’s Gold Medal for having the best academic record in the institute